OXFORD BROOKES UNIVERSITY,

A 1-D Theoretical
Performance Analysis
and Optimisation of a
Bespoke Formula
Student V-Twin Engine
Masters by Research Thesis

Ben Bradshaw

March 2019
Abstract
This paper examines the engine intended to power a SAE Formula Student car for the Oxford
Brookes Racing Formula Student Team. The engine is an amalgamation of multiple students’
research work over a number of years. Therefore, this projects main aim was validation of the
engine setup and specifications resulting from previous work, and to find any potential
theoretical optimisations that can be implemented before the engine is run for real on a
Dynamometer.

The work analyses the engine's power performance and component design, as well as exploring
the major parameters involved in race-engine design and development. The analysis was
carried out via a 1-D simulation model constructed in GT-Suite. This was combined with
combustion and dynamics models constructed in Excel, in addition to component and manifold
analysis in SolidWorks and CATIA. The objectives of the work were –

- Construct a combustion and dynamics model in Excel to analyse the current engine setup.
- Identify areas of the engine design for improvement
- Validate the previous GT-Power model and construct a new GT-Power model based on the
latest version of the GT-Suite software.
- Validate the findings in the Excel models using the GT-Power model.
- Analyse the current engine performance using all the models.
- Propose new hardware for the engine if needed.

The bespoke V-Twin engine development showed a 97-98 Brake Horse Power (BHP)
theoretical Performance limit. However, the starting engine performance was 70 BHP and the
power output was found to be unstable and unpredictable. The analyses of the internal
components found that they were optimal for the power performance expected from the engine,
with the exception that the current design of the intake and exhaust manifolds were found to
limit performance, with the intake plenum found to be too small. The valve timing was also
found to limit the power performance, with the timing requiring changing in order to optimise
performance.

The validation, and the optimising recommended in this project, indicate that the engine’s
theoretical power performance could be improved from 70 BHP to 90 BHP. The engine with
the improvements recommended in this project can now be mounted on a Dynamometer and
the final stages of the engine’s development can be completed in order to approve it for racing.

1
Acknowledgements
The Author would like to thank Dr. Changho Yang, for his continued support and mentoring
through the project.

Thanks also go to Dr Fabrizio Bonatesta, for help and support on the project.

Thanks also go to Oxford Brookes University students, Luke Mortlock, Patrick Orchard and
Tom Rogers.

2
Abbreviations

- AFR Air Fuel Ratio

- BDC Bottom Dead Centre
- BHP Brake Horse Power
- CA Crank Angle
- CAD Computer Aided Design
- CFD Computational Fluid Dynamics
- CR Compression Ratio
- CoG Centre of Gravity
- CAI Compression Auto Ignition
- EGR Exhaust Gas Recirculation
- EOC End of Combustion
- EVO Exhaust Valve Opening
- EVC Exhaust Valve Closing
- FS Formula Student
- FSAE Formula Society of Automotive Engineers
- GDI Gasoline Direct Injection
- HRR Heat Release Rate
- MEP Mean Effective Pressure
- BMEP Brake Mean Effective Pressure
- IMEP Indicated Mean Effective Pressure
- MPS Mean Piston Speed (MPS)
- IVO Inlet Valve Opening
- IVC Inlet Valve Closing
- IPS Instantaneous Piston Speed
- ID Inlet Diameter
- LHV Lower Heating Value
- MFB Mass Faction Burned
- MAPO Maximum Amplitude of Pressure Oscillations
- NA Naturally Aspirated
- OBR Oxford Brookes Racing
- PFI Port Fuel Injection
- SFC Specific Fuel Consumption
- BSFC Brake Specific Fuel consumption
- SAE Society of Automotive Engineers
- SI Spark Ignition
- SOC Start of Combustion
- ST Spark Timing
- TDC Top Dead Centre
- VL Valve Lift
- VO Valve Overlap

3
Contents
1. Introduction .............................................................................................. 9
2. Literature Review ................................................................................... 14
2.1. Engine performance parameters .......................................................... 14
2.2. Combustion model literature review ................................................... 17
2.3. Dynamics model literature review ...................................................... 24
2.4. Manifold design and optimisation....................................................... 34
2.5. Engine simulation literature review .................................................... 42
2.6. V-twin engine performance and calibration........................................ 44
2.7. CFD literature review .......................................................................... 49
3. Methodology .......................................................................................... 53
3.1. Combustion ......................................................................................... 53
3.2. Dynamics ............................................................................................. 60
3.3. GT – Power model .............................................................................. 66
3.4. Valvetrain methodology ...................................................................... 70
3.5. Manifold methodology ........................................................................ 71
4. Results and discussion ............................................................................ 75
4.1. Combustion model .............................................................................. 75
4.2. Dynamics model data and discussion ................................................. 81
4.3. Manifold design and analysis .............................................................. 99
4.3. V-twin engine performance and calibration...................................... 106
4.4. GT-Power model results. .................................................................. 110
5. Conclusions .......................................................................................... 117
6. References ............................................................................................ 119
1. Bibliography ......................................................................................... 128
Appendix ...................................................................................................... 134
Engineering drawings .................................... Error! Bookmark not defined.

4
Table of Figures

Figure 1: Cross section view of the V-Twin engine from the CATIA model. ...................................... 10
Figure 2: Original Engine Model Layout in GT-Power V6.1 ............................................................... 11
Figure 3: Final Power curve of the previous work carried out by Oxford Brookes students................ 11
Figure 4: Mass Fraction Burned Vs In-cylinder pressure trace compared to crank angle. ................... 21
Figure 5: Engine cylinder schematic and breakdown representation of cylinder volume parameters [31]
.............................................................................................................................................................. 24
Figure 6: Diagram of both the reciprocating and rotating masses and the gas force pressure showing the
forces acting on the engine components. .............................................................................................. 27
Figure 7: Diagram indicating the centre of gravity of the con-rod. ...................................................... 28
Figure 8: Diagram of the rotating mass and balancing mass. ............................................................... 30
Figure 9: Representation of forces acting on the individual engine internal components [33]............. 31
Figure 10: The space where an intake is permitted to be located according to the SAE FS rules [47] 34
Figure 11: Comparison of the similarities between a Helmholtz resonator and a vibration absorber .. 35
Figure 12: Harmonic orders .................................................................................................................. 40
Figure 13: A comparison of Maximum amplitude of pressure oscillations over knocking cycles, [12].
.............................................................................................................................................................. 45
Figure 14: A comparison of knocking and non-knocking pressure traces, used to identify knocking
engine cycles [1]. .................................................................................................................................. 46
Figure 15: A shows irregularly distributed mesh, b shows structured distributed mesh and c shows
partially structured mesh, [92]. ............................................................................................................. 50
Figure 16: a is a first order tetrahedral element with 4 nodes, b is a degenerate first order tetrahedral
element with 8 nodes, c is a second order tetrahedral element with 10 nodes, d is a degenerate second
order tetrahedral with 20 nodes, [6]. ..................................................................................................... 51
Figure 17: Otto cycle Pressure Volume profile .................................................................................... 53
Figure 18: A representation of the combustion model constructed in Excel. ....................................... 55
Figure 19: Volume displacement profile from Excel combustion model. ............................................ 56
Figure 20: Pressure comparison of the Otto cycle and Wiebe function. ............................................... 56
Figure 21: Temperature comparison of Otto cycle and Wiebe functions. ............................................ 57
Figure 22: Pressure Vs Volume graphs for Wiebe function and Otto Cycle. ....................................... 57
Figure 23: Shaft work for both Otto cycle and Wiebe function. ........................................................... 58
Figure 24: Torque per crank angle for the Otto cycle and Wiebe function. ......................................... 58
Figure 25: dQ/dθ profile. ...................................................................................................................... 59
Figure 26: MFB profile for the Wiebe function. ................................................................................... 59
Figure 27: A representation of the dynamics model constructed in Excel............................................ 61
Figure 28: Piston displacement. ............................................................................................................ 61
Figure 29: piston velocity. .................................................................................................................... 62
Figure 30: Piston acceleration. .............................................................................................................. 62
Figure 31: Comparison of the primary and secondary forces. .............................................................. 63
Figure 32: Combination of the primary and secondary forces due to the masses. ................................ 63
Figure 33: Gas force applied onto the piston. ....................................................................................... 64
Figure 34: Analysis of the effects of force on same phase firing orders............................................... 64
Figure 35: Analysis of the effects of force on out of phase firing orders. ............................................ 65
Figure 36: Analysis of the total force over the RPM range. ................................................................. 65
Figure 37: The original engine model setup used in version GT-Power 6.1. The work was done carried
out at Oxford Brookes Universities by the students there..................................................................... 67

5
Figure 38: Original model of the engine used to analyse the changes proposed for the engine. In GT-
Power V 7.1. This was the original starting point of the project. ......................................................... 67
Figure 39: GT-Power model development flow chart .......................................................................... 68
Figure 40: The final setup for the engine model in version GT-Power V2016. ................................... 68
Figure 41: A representation the sweep function on GT-Power............................................................. 69
Figure 42: Comparison of Exhaust and Intake valve opening profiles Vs Displacement Volume,
showing original valve timings and areas of importance such as valve overlap................................... 70
Figure 43: CAD of the AT-Power FSAE Throttle Part number from the FS servers FS08-02-520 (FS
CAD) [103] ........................................................................................................................................... 71
Figure 44: Pressure trace from the inlet runner simulated in GT-Power. ............................................. 73
Figure 45: Heat release per crank angle ................................................................................................ 75
Figure 46: Mass Fraction Burned Vs Crank angle. ............................................................................... 76
Figure 47: Displacement per cylinder of the V-Twin including clearance volume .............................. 76
Figure 48: Temperature comparison of the Wiebe and Otto cycles calculated using the Excel template.
.............................................................................................................................................................. 78
Figure 49: Pressure comparison of the Wiebe and Otto cycles, calculated using the Excel Template. 79
Figure 50: Piston velocities by different con rod lengths. .................................................................... 81
Figure 51: Piston velocities by different crank shaft offset lengths. ..................................................... 82
Figure 52: Piston acceleration by different con rod lengths. ................................................................ 82
Figure 53: Piston acceleration by different crank shaft offset lengths. ................................................. 83
Figure 54: Analysis of the con-rods centre of mass using CATIA R5V24........................................... 85
Figure 55: The two sections of con-rod for each assembly of the reciprocating and the rotating masses.
.............................................................................................................................................................. 85
Figure 56: CATIA Piston Assembly to analyse the mass. .................................................................... 86
Figure 57: Piston velocity graph from Excel model. ............................................................................ 87
Figure 58: Analysis of both pistons velocities. ..................................................................................... 88
Figure 59: Analysis of piston acceleration which is a key parameter when analysing inertial component
loads due to mass. ................................................................................................................................. 88
Figure 60: Primary and Secondary reciprocating forces ....................................................................... 89
Figure 61: Total reciprocating forces over a full 4 Stroke cycle........................................................... 90
Figure 62: Crank offset above the crank shaft axis. .............................................................................. 91
Figure 63: Rotating mass assembly above the axis of the crank shaft. ................................................. 91
Figure 64: Total mass of rotating assembly with all the components. .................................................. 92
Figure 65: Analysis of counterweight in CATIA V5R20. .................................................................... 93
Figure 66: Finished rotating assembly with counter weight added in................................................... 94
Figure 67: Torque inertia of the combined reciprocating masses over the 720 degree engine. ............ 95
Figure 68: Pressure plot from GT-Post and plot of Gas force from Excel dynamics model. ............... 96
Figure 69: Combined Gas forces and inertia of reciprocating masses from the Dynamics model. ...... 97
Figure 70: A shows out of phase firing order loads. B shows same phase firing loads. ....................... 98
Figure 71: CAD model of the original intake manifold ........................................................................ 99
Figure 72: Pressure trace from the inlet runner simulated in GT-Power. ........................................... 101
Figure 73: Runner angle analysis and curvature. ................................................................................ 102
Figure 74: Length reducing expansion representation of the runner. ................................................. 102
Figure 75: Flow analysis of exhaust pressure distribution using CFD, [75]. ...................................... 103
Figure 76: Analysis of flow within the current and proposed exhaust system, [75]. .......................... 104
Figure 77: Side by side comparison of the pressure within the current exhaust and the original, [75].
............................................................................................................................................................ 104

6
Figure 78: Side by side comparison of the velocity within the current exhaust and the original, [75].
............................................................................................................................................................ 105
Figure 79: CAD representation of what a potential intake manifold could look like with the proposed
modifications......................................................................................... Error! Bookmark not defined.
Figure 80: Max Horse Power Valve sweep ........................................................................................ 106
Figure 81: Max Torque Valve sweep. ................................................................................................. 107
Figure 82: Average Horse Power Valve sweep .................................................................................. 107
Figure 83: Average Torque Valve Sweep ........................................................................................... 108
Figure 84: Detailed intake valve Sweep ............................................................................................. 109
Figure 85: Detailed exhaust valve Sweep ........................................................................................... 109
Figure 86: Original vs final GT-Power model in-cylinder pressure data............................................ 111
Figure 87: Original and Final GT-Power model Pressure vs Volume. ............................................... 112
Figure 88: Original GT-Power Mass Fraction Burned trace. .............................................................. 113
Figure 89: Final GT-Power model Mass Fraction Burned.................................................................. 113
Figure 90: Original GT-Power models power performance curve. .................................................... 114
Figure 91: Final GT-Power models Power Performance curve. ......................................................... 114
Figure 92: Power Curve Comparison.................................................................................................. 115
Figure 93: Ambient Temperature setup. ............................................................................................. 134
Figure 94: Initial fluid state................................................................................................................. 134
Figure 95: Wiebe model parameters for the GT-Power model ........................................................... 135
Figure 96: Combustion model settings. .............................................................................................. 135
Figure 97: Combustion chamber geometry .stl files. .......................................................................... 136
Figure 98: Combustion chamber geometry input graph ..................................................................... 136
Figure 99: Engine Crank Train model. ............................................................................................... 136
Figure 100: Engine Friction model. .................................................................................................... 137
Figure 101: Engine internals’ parameters. .......................................................................................... 138
Figure 102: Rotating inertia of the internals. ...................................................................................... 138
Figure 103: Max Torque optimiser sweep. ......................................................................................... 139
Figure 104: Exhaust length optimiser sweep ...................................................................................... 139
Figure 105: Intake runner length optimiser sweep.............................................................................. 140
Figure 106: Exhaust valve timing optimiser sweep. ........................................................................... 140
Figure 107: Intake Valve timing optimiser sweep .............................................................................. 141
Figure 108: Plenum volume optimiser sweep ..................................................................................... 141
Figure 109: Fuel parameters [22, 23] .................................................................................................. 143

7
 Chapter 1. Introduction

8
 1. Introduction
A bespoke V-twin Spark Ignition (SI) engine has been developed by Students in the
Mechanical Engineering and Mathematical Sciences department at Oxford Brookes University.
The engine has been in development since 2006 for Oxford Brookes Racing (OBR) Formula
Student (FS) team that competes in the Society of Automotive Engineers (SAE) internal
Formula Student events alongside university teams from around the world. The engine is a
600CC Naturally Aspirated (NA) SI engine that is expected to be in the car by 2018/19.

The engine was started by Oxford Brookes University students as a theoretical project in 2005.
The engine is based on 2 cylinders of a Cosworth DFV Formula 1 V8. Although the engine
was designed and built from scratch, the first students used the DFV engine as a base for
research into the area of small capacity engines that would be competitive in FS events. The
engine is part of the first full engine and powertrain package designed and built entirely by
students.

The project will explore the design of the engine's components and analyse the major
parameters involved in race engine design and development. The aim and objectives of this
project are –

- The aim of the project is to validate the current engine set up and components and to optimise
the engine where possible

Objectives –
- Construct a combustion and dynamics model in Excel to analyse the current engine setup.
- Identify areas of the engine design for improvement
- Validate the previous GT-Power model and construct a new GT-Power model based on the
latest version of the GT-Suite software.
- Validate the findings in the Excel models using the GT-Power model.
- Analyse the current engine performance using all the models.
- Propose new hardware for the engine if needed.

The project will look into the design and optimisation of inlet and exhaust systems and create
Excel spread sheets to determine the key factors for cam design, and inlet and exhaust tuning.

This study seeks to optimise the engine where real-time data from testing the engine is not
available. Therefore, the Heat Release Rate (HRR) and the Wiebe function were used to predict
the engine’s combustion properties in conjunction with an Excel spreadsheet.

The project was subject to limitations, one of which was that not all parts of the engine and
drivetrain had been built. For example, it was not possible to calibrate the engine as the gearbox
had not been built. The project identified areas of the engine where changes would lead to
improvement in engine performance. A cross section of the original engine can be seen in
figure 1 and the specifications of the engine at the start of the project can be seen in Table 1.

9
 Figure 1: Cross section view of the V-Twin engine from the CATIA model.

10
 Figure 2: Original Engine Model Layout in GT-Power V6.1

The model seen in Figure 2: Original Engine Model Layout in GT-Power V6.1 is the
reconstruction of the original model in the latest version of the GT-Power software, as the
original software was out of date and was not cross compatible.

Figure 3: Power curve of the engine at the start of the project.

As can be seen in Figure 3: Power curve of the engine at the start of the project, the previous
work on the engine had a smooth power curve with peak power performance of 112 Brake
Horse Power (BHP) at 11,000 RPM. This engine performance and power curve was not
achievable by the theoretical analysis and from the restrictor required by the FS regulations.

11
 BASE ENGINE SPECIFICATIONS
BASE ENGINE PARAMETERS
COMPRESSION RATIO (STATIC) RATIO 12.45
COMPRESSION RATIO (DYNAMIC, 0.5MM ROD STRETCH) RATIO 14.30
BORE MM 92.00
SWEPT VOLUME CC 600.00
STROKE MM 45.13
IDLE RPM RPM 4000
PEAK TORQUE RPM RPM 9000
MAX RPM RPM 12500
ANGLE OF CYLINDERS (V-ANGLE) DEG 75.00
CON-ROD
CON-ROD LENGTH MM 103.00
CON-ROD BIG END WIDTH MM 20.00
VALVE AND PORTS
INLET VALVE DIAMETER MM 38.12
INLET VALVE STEM DIAMETER MM 7.00
EXHAUST VALVE DIAMETER MM 31.85
EXHAUST VALVE STEM DIAMETER MM 7.00
INLET PORT DIAMETER (AT VENTURE THROAT) MM 27.50
EXHAUST PORT DIAMETER (BEHIND SEAT VALVE) MM 26.00
TAPPET DIAMETER MM 33.764
TAPPET BORE DIAMETER MM 33.65
INLET PORT LENGTH MM 61.00
EXHAUST PORT LENGTH MM 55.00
TOTAL INLET TRACT LENGTH (BELLMOUTH TO VALVE SEAT) MM 326.80
HARMONIC ORDER 3.5
INLET VALVE LIFT MM 10.978
EXHAUST VALVE LIFT MM 9.727
CAMSHAFT
CAMSHAFT BEARING SURFACE DIAMETER MM 25.00
CAM BEARING INTERNAL DIAMETER MM 25.050
CRANKSHAFT
CRANKSHAFT MAIN JOURNAL BEARING MM 49.20
CRANKSHAFT SMALL END JOURNAL DIAMETER MM 42.00
CYLINDER HEAD/BLOCK
CYLINDER HEAD STUD ANGLE DEG 24.00
CYLINDER BLOCK DECK HEIGHT FROM CRANK CR MM 149.264
OTHER
MAIN OIL GALLERY INSIDE DIMETER MM 12.00
CRANK CR TO HALF SPEED STACK CR MM 82.32
MEAN INLET TRACT AIRSPEED AT 12,500RPM M/S 105.00

Table 1: Original Engine Specifications

12
Chapter 2. Literature review

13
 2. Literature Review
In this section the basics of an SI engine will be explored and broken down into their basic
parameters. Although there are many different types of internal combustion engine, the V-Twin
is a 4-stroke, spark ignited, petroleum fuelled (although other spark ignited fuels can be used)
engine that converts heat addition and air expansion into rotary motion via the use of a sliding
piston connected to a rotating crankshaft. The fundamentals of a 4-stroke engine are the Otto
cycle, which can be simplified to the ignition stroke, exhaust stroke, induction stroke and
compression stroke. Each stroke has a different characteristic in terms of pressure, temperature
and air flow in and out of the cylinder. As there are multiple process going on within the engine
that cannot be explained with one expression, each process/action must be broken down into
smaller equations that can describe the different actions, parameters and characteristics.

2.1. Engine performance parameters

There are many ways of assessing engine performance. This section will explore the basics of
some of them in order to further the understanding of engine design and development. The
Indicated Work per cycle allows the engine’s performance to be assessed using a simple
pressure trace. Indicated Work per cycle is given by the equation, [1, 2] –
+𝟑𝟑𝟑𝟑𝟑𝟑

𝑾𝑾𝒊𝒊 = � 𝒑𝒑 𝒙𝒙 𝒅𝒅𝒅𝒅 ≈ − � 𝒑𝒑 × ∆𝑽𝑽

−𝟑𝟑𝟑𝟑𝟑𝟑

Equation 1

The Indicated Power, which depends on engine speed, can then be found using the Indicated
Work using the equation, [2] –
𝑾𝑾𝒄𝒄 × 𝑵𝑵
𝑷𝑷𝒊𝒊 =
𝒏𝒏𝒓𝒓

Equation 2

Mean Effective Pressure (MEP), is the work per unit volume and is an artificial parameter that
cannot be measured, it is independent of engine displacement and therefore allows engines of
varying displacements to be compared more easily, [2, 3] –

𝑊𝑊 𝑊𝑊 𝑤𝑤
𝑀𝑀𝑀𝑀𝑀𝑀 = = =
𝛥𝛥𝛥𝛥 𝑉𝑉𝑑𝑑 𝛥𝛥𝛥𝛥
Equation 3

Or simply –

𝑊𝑊𝑊𝑊𝑊𝑊𝑊𝑊 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 𝑝𝑝𝑝𝑝𝑝𝑝 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐

𝑀𝑀𝑀𝑀𝑀𝑀 =
𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆 𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣
Equation 4

MEP can be broken down into different definitions, Indicated and BMEP are defined by, [2] –

𝑊𝑊𝑖𝑖 𝑊𝑊𝑏𝑏 𝑊𝑊𝑖𝑖 × 𝜂𝜂𝑚𝑚

𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼 = and 𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵 = = = 𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼 × 𝜂𝜂𝑚𝑚
𝛥𝛥𝛥𝛥 𝛥𝛥𝛥𝛥 𝛥𝛥𝛥𝛥
Equation 5

14
The Brake is the Indicated multiplied by the mechanical efficiency. To find the Mechanical
Efficiency, the BMEP can be divided by the Mean Effective Pressure or the Brake Power can
be divided by the Indicated Power, [2, 4] –

𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵

𝜂𝜂𝑚𝑚 = =
𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼
Equation 6

To find the BMEP, the IMEP has to be multiplied by the Mechanical Efficiency

𝑝𝑝 × 𝜂𝜂𝑚𝑚
𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵 = 𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼 × 𝜂𝜂𝑚𝑚 =
𝛥𝛥𝛥𝛥
Equation 7

Air Fuel Ratio (AFR) is the ratio of fuel to air inside the combustion chamber. If oxidants other
than air are used for combustion, the same equation applies. The AFR is given by the equation
[1, 2, 5] –
𝑚𝑚𝑎𝑎
𝐴𝐴𝐴𝐴𝐴𝐴 =
𝑚𝑚𝑓𝑓
Equation 8

The Equivalence Ratio, ϕ, is a variable that gives an indication on how similar the AFR is to
the stoichiometric AFR. The stoichiometric AFR is the amount of air and fuel in the
combustion chamber to required give complete combustion. The Equivalence Ratio is given
by the equation [1, 2, 3, 6] –
𝐴𝐴𝐴𝐴𝐴𝐴𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎
𝜙𝜙 =
𝐴𝐴𝐴𝐴𝐴𝐴𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠ℎ𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖
Equation 9

The Equivalence Ratio can indicate how the engine is running, for example [2] –

ϕ = 1 – Ideal fuelling (stoichiometric combustion)

ϕ > 1 – Rich mixture (too much fuel for stoichiometric combustion)
ϕ < 1 – Lean mixture to the engine (not enough fuel for stoichiometric
combustion)

The engine will be run at equivalence ratio of 1 which is a stoichiometric air-fuel ratio of 14.7:1
for gasoline.

Specific Fuel Consumption (SFC) is the amount of fuel the engine is consuming for a given
output, usually g/kWh, and can be found using the equation [1, 2] –

ṁ𝑓𝑓
𝑆𝑆𝑆𝑆𝑆𝑆 =
Ẇ
Equation 10

Volumetric Efficiency, ηv, shows how much air an engine can intake as a ratio of the total
volume of the engine to the volume of air inducted during the intake stroke, at ambient
conditions. Volumetric efficiency can show how much oxygen an engine has to burn and can
give an indication of the correct amount of fuel needed for optimum performance. As the air

15
density is a factor in determining the ηv, if the density is taken from within the inlet system
then only the flow losses around the valve and through the valve seat will be taken into account
in the ηv. Typically, around 80 to 90% on road cars, it can be higher in race engines where
expected lifespan is far shorter due to the much lower time the engine will be powering the car,
often only a few races depending on the race series, [1, 7].

𝑚𝑚𝑎𝑎 𝑚𝑚̇𝑎𝑎 𝑛𝑛
𝜂𝜂𝑉𝑉 = =
𝜌𝜌𝑎𝑎,𝑖𝑖 𝑉𝑉𝑑𝑑 𝜌𝜌𝑎𝑎 𝑉𝑉𝑑𝑑 𝑁𝑁

Equation 11

Where ηv is the Volumetric Efficiency, ma is the mass of air, Vd is the displacement volume
and ρa,i is the density of the air at the inlet, N is the speed of the engine and n is the number of
crank shaft rotations per complete engine cycle. Inlet density may be taken as atmospheric but
that will mean the losses within the inlet system will also be shown in the overall ηv.

Combustion Efficiency, ηc, indicates how much of the fuel has been burnt during the cycle.
Lean AFRs normally have 95 to 98% ηC whereas rich mixtures have lower Combustion
Efficiencies due to less oxygen available [1, 7].

ℎ𝑅𝑅 (𝑇𝑇𝐴𝐴 ) − ℎ𝑝𝑝 (𝑇𝑇𝐴𝐴 )

𝜂𝜂𝐶𝐶 =
𝑚𝑚𝑓𝑓 𝑄𝑄𝐿𝐿𝐿𝐿𝐿𝐿

Equation 12

Where hR and hp are the reference and pre-chamber enthalpy respectively and TA is the ambient
air temperature, QLHV is the calorific value of the fuel and mf is the mass of fuel.

16
 2.2. Combustion modelling
An in-cylinder pressure trace is a fundamental part of engine analysis and therefore a
combustion model was deemed necessary to give an in-cylinder pressure trace. Another reason
for constructing a combustion model is to help improve the understanding of how engine
simulation software, such as GT-Power, work and identify any potential problems.

There are multiple types of combustion model, each with a different purpose. The different
models have benefits and drawbacks.

Predictive vs Non-Predictive vs Semi-Predictive

A non-predictive combustion model imposes a burn rate based on pre-set combustion
parameters and does not deviate from this known burn rate. Non-predictive models will give a
pre-determined burn rate and aren’t affected by the in-cylinder environment. A non-predictive
combustion model allows quick analysis and predictions of known combustion profiles.
However, a non-predictive model will not show if any changes of engines parameters will lead
to a significant change in the combustion profile. Non-predictive models are not an appropriate
choice of combustion model for simulating variables that directly affect the burn rate of the
combustion event, [8, 9, 10, 11, 12]. It was found by Dr Hasse et al [13] that a non-predictive
model in GT-Power could cut simulation times by more than 50% compared to a predictive
model.

A predictive combustion model is based on the inputs important to the combustion process,
such as temperature, pressure, spark timing, combustion chamber geometry, valve timing and
fuel. Whilst the results are more accurate than from a none-predictive model, a predictive
combustion model is complex and time consuming and relies on parameters that are hard to
predict without real time in-cylinder data in order to calibrate them, [8, 9, 10, 11, 12].

A compromise between the two is a semi-predictive model. Semi-predictive models are

affected by significant variables, and they react to changes in the variables but do not
incorporate models to predict the outcomes, if these variables are unknown. Semi-predictive
models incorporate non-predictive models, such as the Wiebe model, in order to calculate the
combustion but rely on lookups or pre-set responses to calculate the outcomes of these
significant parameters based on the Wiebe function. Semi-predictive models allow for a
compromise between the accuracy and the calculation times but they require a far more
combustion data on the particular area of interest in order to rely on the lookup directory, [8,
9, 10, 11, 12].

Chemical and empirical model

A chemical combustion model uses the atomic structure of the fuels and oxidants to predict the
combustion event. Chemical combustion models tend to be theoretical and are predictive as
they are based on events that are hard to measure accurately. Although a chemical combustion
model can be an accurate predictive model, it cannot be used to analyse a real-time combustion
event. The therefore data needs to be compared to a real-time combustion event of the same
design in order to validate the results. A chemical kinetic combustion model calculates the
concentration of species and global/elementary mechanism heat release. A chemical kinetic
model combines parameters such as flame propagation, turbulent flow fields, diffusion fields,
heat exchange fields, amongst others. The model can calculate concentrations of radicals and
chemical species at any point within the model. A chemical kinetics system is a lot more
complicated to produce than a chemical equilibrium or empirical model, [8, 9, 10, 11].

17
An empirical combustion model uses mathematical methods to analyse the combustion event.
Empirical models can be used in post event analysis. Empirical methods must incorporate a lot
of assumptions to be used as a predictive method but are useful to analyse the combustion event
in real time or using post event data. A chemical equilibrium combustion model typically
assumes harmonious combustion with a single zone and uniform parameters such as
temperature and pressure, [8, 9, 10, 11].

Flame propagation and Combustion phasing

Flame propagation is the analysis of the combustion event by analysing the development of the
flame from kernel, through to flame propagation, to extinguishment. Flame propagation
analysis the physical flame and through the defined steps of the flames development in the
cylinder [14, 15, 16].

Combustion phasing is the analysis of the combustion event by splitting the combustion profile
into defined sections of in-cylinder charge burned and comparing them to the crank angle
location of sections or anchor points. 50%MFB is commonly used as the combustion phase
anchor and is defined as when half the in-cylinder charge is burned, depending on the CA
location of the anchor can give different combustion characteristics and different engine
performance [14, 15, 16].

Flame propagation is more analytical of the actual combustion event and therefore lends itself
to theoretical analysis of the combustion event, while combustion phasing is the analysis of a
combustion profile which lends itself to real-time and post real-time data combustion analysis.
GT-Power has models for both types of combustion event although it was deemed that for this
project it would be more appropriate to carry out combustion analysis, using combustion
phasing, rather than flame propagation analysis. Combustion phasing was deemed appropriate
for the project as it can be analysed using in-cylinder pressure data. Flame propagation requires
the use of either a complex setup with specialist equipment such as see-through combustion
chambers, or a theoretical chemical kinetic model both where deemed too resource heavy and
time consuming. However, the principals of flame propagation were researched and
investigated, so that it was applied to combustion model as well as the combustion phasing.
Flame propagation theory for of combustion models can be seen in Flame Propagation section
of the Appendix [14, 15, 16].

18
 2.2.1. Heat Release Rate
HRR is the rate at which energy, in terms of heat, is released by the fuel when it is burned.
HRR is a critical parameter in terms of engine performance. The HRR inside a cylinder is
typically larger than if the same air/fuel charge was combusted in a constant volume, closed
system, due to the compression of the air inside the cylinder as the piston travels to Top Dead
Centre (TDC). HRR can be calculated using different methods. The apparent HRR assumes
that the change in pressure from compression and expansion for a given change in crank angle
is proportional to the fuel burnt over the same interval and so the apparent HRR is calculated
as the heat released per change in crank angle θ as a function of gamma, if a perfect gas is
assumed. [2, 18] –

𝛿𝛿𝑄𝑄ℎ𝑟𝑟 = [ 𝑑𝑑𝑑𝑑 ] + [ 𝛿𝛿𝛿𝛿 ] + [ 𝛿𝛿𝑄𝑄𝐻𝐻𝐻𝐻 ]

Equation 13

this can be broken down into [2] –

𝑑𝑑𝑄𝑄ℎ𝑟𝑟 𝐶𝐶𝑣𝑣 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑 𝑑𝑑 𝑄𝑄ℎ𝑡𝑡

=� � 𝑝𝑝 + 𝑉𝑉 �� + � 𝑝𝑝 �+� �
𝑑𝑑𝑑𝑑 𝑅𝑅 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑
Equation 14

and further [2, 7] –

𝑑𝑑𝑄𝑄𝑛𝑛𝑛𝑛𝑛𝑛 𝐶𝐶𝑣𝑣 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑 𝐶𝐶𝑣𝑣 𝑑𝑑𝑑𝑑 𝐶𝐶𝑣𝑣 𝑑𝑑𝑑𝑑

=� � 𝑝𝑝 + 𝑉𝑉 �� + � 𝑝𝑝 � = � + 1� 𝑝𝑝 + 𝑉𝑉
𝑑𝑑𝑑𝑑 𝑅𝑅 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑 𝑅𝑅 𝑑𝑑𝑑𝑑 𝑅𝑅 𝑑𝑑𝑑𝑑
Equation 15

and finally [2, 7] –

𝑑𝑑𝑄𝑄𝑛𝑛𝑛𝑛𝑛𝑛 𝛾𝛾 𝑑𝑑𝑑𝑑 1 𝑑𝑑𝑑𝑑

= � × 𝑝𝑝 � + � × 𝑉𝑉 �
𝑑𝑑𝑑𝑑 𝛾𝛾 − 1 𝑑𝑑𝑑𝑑 𝛾𝛾 − 1 𝑑𝑑𝑑𝑑
Equation 16

where γ is the ratio of specific heats (typically 1.3 ~ 1.4 ).

One method is the single zone. The single zone equation is so-called as it uses an averaged
value for gamma from the two zones, the two zones being unburnt and burnt. It is worked out
using the equation [2, 7] –

𝑄𝑄𝐿𝐿𝐿𝐿𝐿𝐿× 𝑚𝑚𝑐𝑐 ℎ × 𝑉𝑉𝑐𝑐𝑐𝑐

� � × 𝑑𝑑𝑑𝑑𝑏𝑏 − 𝑚𝑚𝑐𝑐 × 𝐶𝐶𝐶𝐶 × 𝑑𝑑𝑑𝑑 = 𝑝𝑝 × 𝑑𝑑𝑑𝑑 + 𝐴𝐴 × ℎ� × (𝑇𝑇 − 𝑇𝑇𝑤𝑤 ) + 𝑑𝑑𝑑𝑑
𝐴𝐴 �
𝑅𝑅
𝐹𝐹 + 1 𝑀𝑀 × 𝑇𝑇𝑤𝑤
Equation 17

where Cv is the specific heat at constant volume, xb is the mass fraction burned, mc is the mass
of the charge, T is temperature, p is pressure, V is volume, ℎ� is the coefficient of average heat
transfer, Tw is temperature of the wall, A/F is the air fuel ratio, QHLV is the lower heat value, A
is the combustion chamber surface area.

A further equation can be used where the specific heats from the two zones are not averaged.
This is a much more complicated equation [2, 18] –

19
 𝑄𝑄𝐿𝐿𝐿𝐿𝐿𝐿× 𝑚𝑚𝑐𝑐
� + 𝑚𝑚𝑐𝑐 × 𝐶𝐶𝐶𝐶𝑢𝑢 × 𝑇𝑇𝑢𝑢 − 𝑚𝑚𝑐𝑐 × 𝐶𝐶𝐶𝐶𝑏𝑏 × 𝑇𝑇𝑏𝑏 � × 𝑑𝑑𝑑𝑑𝑏𝑏 − 𝑚𝑚𝑐𝑐 × (1 − 𝑥𝑥𝑏𝑏 ) × 𝐶𝐶𝐶𝐶𝑢𝑢 × 𝑑𝑑𝑑𝑑𝑢𝑢
𝐴𝐴
𝐹𝐹 + 1
− 𝑚𝑚𝑐𝑐 × 𝑥𝑥𝑏𝑏 × 𝐶𝐶𝐶𝐶𝑏𝑏 × 𝑑𝑑𝑑𝑑𝑏𝑏 − 𝑝𝑝 × 𝑑𝑑𝑑𝑑𝑢𝑢 − 𝑝𝑝 × 𝑑𝑑𝑑𝑑𝑏𝑏
ℎ𝑢𝑢 × 𝑉𝑉𝑐𝑐𝑐𝑐
= 𝐴𝐴𝑢𝑢 × ℎ�𝑢𝑢 × (𝑇𝑇𝑢𝑢 − 𝑇𝑇𝑤𝑤 ) + 𝐴𝐴𝑏𝑏 × ℎ�𝑏𝑏 × (𝑇𝑇𝑏𝑏 − 𝑇𝑇𝑤𝑤 ) + 𝑑𝑑𝑑𝑑
𝑅𝑅�𝑢𝑢
𝑀𝑀𝑢𝑢 × 𝑇𝑇𝑤𝑤
Equation 18

where all the symbols mean the same as for the single zone, but the suffix u means unburnt and
the suffix b means burnt. The integral of HRR can be represented as MFB. The MFB represents
the amount of the fuel/air charge that has been burned per cycle. [18].

20
 2.2.2. Mass Fraction Burned
Mass Fracture Burn (MFB) is a common combustion phasing indicator to determine the amount
of fuel and air mixture that has been burned for a given crank angle, [19].

Figure 4: Mass Fraction Burned Vs In-cylinder pressure trace compared to crank angle.

Figure 4: Mass Fraction Burned Vs In-cylinder pressure trace compared to crank angle shows
an in-cylinder pressure trace compared to a MFB profile, both are represented on a graph
compared to crank angle and will be discussed later in the project.

21
MFB is the fraction of fuel burnt during combustion in relation to crank angle θ. One way of
working out MFB is using the Wiebe Function, which is calculated using [19, 20]–

𝜃𝜃 − 𝜃𝜃0 𝑚𝑚+1
𝑥𝑥𝑏𝑏 = 1 − exp[ −𝑎𝑎 × � � ]
𝛥𝛥𝛥𝛥
Equation 19

where xb is the MFB, θ0 is the crank angle at the start of combustion, Δθ is the crank angle
duration of combustion, m is the form factor, a is the efficiency parameter, and b is the
amplitude correction factor, [18, 21]. Typically, a and m for an SI engine are a=5 and m=3, [1,
7, 41, 102].

μb (i) ∑𝑖𝑖𝑘𝑘=0 ∆𝑝𝑝𝑐𝑐 (𝑘𝑘)

xb = = 𝑛𝑛
μb (total) ∑𝑘𝑘=0 ∆𝑝𝑝𝑐𝑐 (𝑘𝑘)
Equation 20

Where n is the total number of crank angles. The MFB is the change in pressure within the
cylinder in proportion to the mass of charge that is burnt. Another way of calculating MFB is
to use pressure differentials form experimental data as seen in equation 20.

The MFB is commonly determined, in SI engines, by analysing the burn rate. This technique
was developed by Rassweiler and Withrow [26], in 1938, and is considered accurate and
computationally efficient, [2, 19] –

𝛥𝛥𝛥𝛥 = 𝛥𝛥𝑝𝑝𝑐𝑐 + 𝛥𝛥𝑝𝑝𝑣𝑣 = 𝑝𝑝1+1 − 𝑝𝑝1

where Δpc is the change in pressure due to combustion and Δpv is the change of pressure due
to the change in volume. As the crank moves through increments of θi through to θi + 1 so will
the volume and the pressure move through increments of i thought to i + 1 [1, 2, 4, 19] –

vi γ
Δpc (i) = pi+1 − pi � �
vi+1
Equation 21

Hence the equation for change in pressure due to combustion (Δpc) can be formed –

𝑣𝑣𝑖𝑖 𝛾𝛾
𝛥𝛥𝑝𝑝𝑣𝑣 = 𝑝𝑝𝑖𝑖+1,𝑣𝑣 − 𝑝𝑝𝑖𝑖 = 𝑝𝑝𝑖𝑖 �� � − 1�
𝑣𝑣𝑖𝑖+1
Equation 22

Once HRR has been calculated, the data can be used to calculate the efficiency of the
combustion. This is done by comparing the total heat released in combustion with the heat that
was available from the fuel using the equation, [18] -

𝑄𝑄
𝜂𝜂𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 =
𝑚𝑚
�𝐴𝐴 𝑐𝑐 � × 𝑄𝑄𝐿𝐿𝐿𝐿𝐿𝐿
𝐹𝐹 + 1
Equation 23

22
where A is the surface area of the combustion chamber, mc is the mass of the charge, Q is heat
and QHLV is the lower heat value.

The MFB curve can be split into three distinct points; 10%, 50% and 90% MFB, to highlight
the combustion phasing [1, 17, 18]:

Start of combustion (10% of MFB) - There is a delay from spark ignition to the start of
combustion that does not add much in terms of work done by the piston. MFB 0 – 10% is
known as early flame development. The spark timing, and the subsequent ignition delay
interval, have a significant effect on the shape and location of the MFB curve for a given Profile
calibration. Therefore, MFB10 is a useful anchor point for comparison of combustion events
and the first 10% is not used in the analysis. Ideally 0 – 100% MFB would be used for
comparison but it is too difficult to experimentally define these locations on a MFB curve while
MFB 10 allows for more accurate comparison of different MFB curves, [1, 17, 18, 22, 23, 24].

Midpoint (50% MFB) - The midpoint of combustion or 50% MFB is used to position the heat
release curve and can give good indication of the efficiency of combustion by comparing it to
TDC of the engine [1, 17, 18]. MFB50 is used as an anchor point to analyse the peak pressure
during combustion as it allows for the combustion profile to be moved via a single reference
point. MFB50 can be used as a good indicator of the point of maximum heat release, although
it is not always accurate, [25, 22, 23, 24].

End of combustion ( 90% MFB) - The last 10% of combustion again adds little to no work
to the piston as there is almost no heat given off from the charge as the majority of the fuel/air
mixture has been burnt. It is hard to estimate when Flame extinguishment happens so again
90% allows more accurate comparison of MFB curves. Again, MFB90 is a useful anchor point
in for comparison of combustion [1, 17, 18]. The coefficients; m, the form factor, a, the
efficiency parameter, and b, the amplitude correction factor, of MFB are adjusted using the
10%, 50% and 90% combustion phasing points to get a more realistic combustion model, [18,
21, 22, 23, 24]. 90% MFB is used as there is often a ‘long burn out phase’ when analysing the
real-time data, which is associated with the bad signal to noise ratio as found by [13].

Combustion duration (10-90% MFB) - The combustion duration MFB10-90 is an important

parameter in engine development due to the amount of information that can be gathered from
it, such as the speed of combustion and how that relates to the speed and movement of the
piston, [22, 23, 24].

23
 2.3. Dynamics model literature review
This section will look into the basic expressions and equations that describe component motion
by displacement, velocity and acceleration.

Figure 5: Engine cylinder schematic and breakdown representation of cylinder volume

parameters [31, 110, 111]

The Compression Ratio is the ratio of the Total Volume to the Clearance Volume. The
Clearance Volume is the volume at TDC and is sometimes referred to as the Combustion
Chamber. The Swept Volume is the total volume the piston sweeps, and fluctuates from no
volume at TDC to the max engine displacement at Bottom Dead Centre (BDC), [38, 39]. –

( 𝑉𝑉𝑑𝑑 + 𝑉𝑉𝑐𝑐 )
𝑟𝑟𝑐𝑐 =
𝑉𝑉𝑐𝑐
Equation 24

Due to the oscillations of the piston going up and down the stroke, a term can be formed by
combining the total stroke length and the speed of the engine, this gives an indication of how
many times, and how far, the piston has moved along the stroke per unit of time.

Piston displacement
The cylinder volume is the given by the equation –

𝜋𝜋 × 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 2
𝑉𝑉 = × 𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙ℎ 𝑜𝑜𝑜𝑜 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐
4
Equation 25

The Bore is the diameter and the Stoke is the length of the cylinder so therefore the Swept
volume is given by [38, 39] –

24
 𝜋𝜋
𝑉𝑉𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 = × 𝐵𝐵 2 × 𝑆𝑆 = 𝑉𝑉𝑇𝑇𝑇𝑇𝑇𝑇 + 𝑉𝑉𝐵𝐵𝐵𝐵𝐵𝐵
4
Equation 26

The Cylinder Volume is the sum of the Clearance Volume Vc and the Displacement Volume
Vd as a factor of θ, [1, 40]. The Clearance Volume, Vc, is defined by the area of the cylinder
and the combustion chamber when the piston is at Top Dead Centre (TDC).

𝜋𝜋 × 𝐵𝐵 2
𝑉𝑉 = 𝑉𝑉𝑐𝑐 + × 𝑋𝑋
4
Equation 27

where X is the piston distance from TDC [1, 32]. To analyse the piston characteristics, the
distance between the piston and the centre of axis of the crank shaft, X, has to be calculated to
give the distance at any point in the stroke. Furthermore, the velocity and acceleration can be
calculated by the derivative of the original distance calculation. [3, 5, 38, 39]. X can be found
by the differential equation [1, 39, 41, 42] –
1
𝑠𝑠 𝑠𝑠 2 2
𝑋𝑋 = cos 𝜃𝜃 + 𝐿𝐿 �1 − � sin 𝜃𝜃� �
2 2𝐿𝐿
Equation 28

where X is the piston position, s is the length of stroke and is equal to twice the crankshaft
1
𝑠𝑠 2 2
offset r, cos 𝜃𝜃 + 𝐿𝐿 is the primary force and �1 − �2𝐿𝐿 sin 𝜃𝜃� � is the secondary force term. L is
the con-rod length

Piston velocity
The second differential of the piston distance equation gives Velocity [38, 39, 41, 42] –

𝑠𝑠 𝜀𝜀
𝑋𝑋̇ = − 𝜔𝜔 �sin 𝜃𝜃 + sin 2𝜃𝜃�
2 2
Equation 29

where 𝑋𝑋̇ is the approximation of the piston velocity, ω is the angular velocity and equates to
the change in crank angle over change in time, dθ/dt, and ε is the con-rod-to-crankshaft offset
ratio, and is equal to s/2L. The piston velocity will give an indication of flow velocity in and
out of the engine, and the applied forces on the components, such as the con-rod and the journal
bearings. It is an important factor in the analysis and development of the manifolds and the
valve seats.

The equation is a function of time t and as θ=ωt. It will be preferable to present the equations
as a function of crank angle θ, [1, 7] –

𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑

𝑋𝑋̇ = 𝑆𝑆𝑝𝑝 = = = 2𝜋𝜋𝜋𝜋
𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑
Equation 30

therefore [38, 39] –

25
 𝑑𝑑𝑑𝑑 1 1 𝑎𝑎 sin 𝜃𝜃 cos 𝜃𝜃
= −𝑎𝑎 sin 𝜃𝜃 + (𝐿𝐿2 − 𝑎𝑎2 sin2 𝜃𝜃)2 (−2𝑎𝑎2 sin 𝜃𝜃 cos 𝜃𝜃) = −𝑎𝑎 sin 𝜃𝜃 −
𝑑𝑑𝑑𝑑 2 √𝐿𝐿2 − sin2 𝜃𝜃
Equation 31

therefore [38, 39] –

𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑 𝑎𝑎 sin 𝜃𝜃 cos 𝜃𝜃 cos 𝜃𝜃

�𝑆𝑆𝑝𝑝 � = = 2𝜋𝜋𝜋𝜋 = 2𝜋𝜋𝜋𝜋 �𝑎𝑎 sin 𝜃𝜃 − � = 2𝜋𝜋𝜋𝜋𝜋𝜋 sin 𝜃𝜃 �1 + �
𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑 √𝐿𝐿2 − sin2 𝜃𝜃 √𝐿𝐿2 − sin2 𝜃𝜃
Equation 32

which leads to the piston speed as a function of crank angle θ [38, 39] –

𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐
�𝑆𝑆𝑝𝑝 � = ( 𝜋𝜋 × 𝑆𝑆 × 𝑁𝑁 ) × 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 �1 + �
√𝑅𝑅 2 − 𝑠𝑠𝑠𝑠𝑠𝑠2 𝜃𝜃
Equation 33

To find Instantaneous Piston Speed, the ratio must be used [38, 39] –

𝑆𝑆𝑝𝑝 𝜋𝜋 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐
= 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 �1 + �
𝑆𝑆𝑝𝑝̅ 2 √𝑅𝑅 2 − 𝑠𝑠𝑠𝑠𝑠𝑠2 𝜃𝜃
Equation 34

The Instantaneous Piston Speed (IPS) allows the speed of the piston to be analysed through the
cycle and also allows for the maximum speed of the piston to be found, which gives the air
flow velocity for the plenum analysis.

Piston acceleration
The third differential gives Acceleration [1, 39, 41, 42] –

𝑠𝑠
𝑋𝑋̈ = − 𝜔𝜔2 (cos 𝜃𝜃 + 𝜀𝜀 cos 2𝜃𝜃)
2
Equation 35

Where 𝑋𝑋̈ is the approximation of the piston acceleration.

26
Engine balancing
Engine balancing is a critical part of engine development because the internal components of
the engine are subjected to large stress loads from the high inertia and gas loads applied to
them. If rotating masses are not balanced then undue stress will be exerted on the engine, with
the potential to cause fatigue and drastically reduce the life of the engine components. A
balanced engine will not only run more smoothly, but potentially will be more efficient as there
will be less force opposing the power stroke. In a piston-driven internal combustion engine the
force from the pressure within the cylinder, created by the fuel combusting, is converted into a
rotational motion by the crank slider mechanism [1, 7, 43]. To understand the engine dynamics
and investigate the engine balancing the dynamics can be split into different sections [38, 39],

• Kinematics
• Pressure forces
• Momentum from both torque and inertia
• Balancing multiple cylinders

Analysis of engine dynamics and balancing can be split into 2 different sections; rotating and
reciprocating masses. The motions of these masses cause inertia forces and vibrations within
the engine [38, 39].

Figure 6: Diagram of both the reciprocating and rotating masses and the gas force pressure showing the
forces acting on the engine components.

P, as seen in Figure 6: Diagram of both the reciprocating and rotating masses and the gas force
pressure showing the forces acting on the engine components., is the pressure force acting on

27
the piston and it is internal so there should be minimal vibrations associated with it. Mass A is
the rotating mass. Mass B is the reciprocating mass and must have the piston mass added to it.

2.3.1. Analysis of reciprocating and rotating masses

At the start of the development of the engine a lot of critical parameters are not known, such
as the mass of the components, the Centre of Gravity (CoG) and the inertia forces. Due to the
complexity of its motion and because within a piston there is a reciprocating mass and a rotating
mass connected by the connecting rod, the con-rod has to be split into two components for
anaylsis, one mass connected to the gudgeon pin and the other to the big end/crank pin [1, 7,
38, 39, 43].

Figure 7: Diagram indicating the centre of gravity of the con-rod.

The total mass of the con-rod has been split into two masses, Mass A and Mass B, shown in
Figure 7: Diagram indicating the centre of gravity of the con-rod.. This figure shows the CoG
of the con-rod. Mass A incorporates the mass of the con-rod acting on the crank pin while Mass
B is the mass of the con-rod acting on the gudgeon pin and the mass of the piston assembly [1,
7, 38, 39, 42, 44, 45].

𝐿𝐿𝑏𝑏 𝐿𝐿𝑎𝑎
𝑚𝑚𝐴𝐴 = 𝑚𝑚𝑐𝑐 𝐚𝐚𝐚𝐚𝐚𝐚 𝑚𝑚𝐵𝐵 = 𝑚𝑚𝑐𝑐 + 𝑚𝑚𝑝𝑝
𝐿𝐿𝑎𝑎 + 𝐿𝐿𝑏𝑏 𝐿𝐿𝑎𝑎 + 𝐿𝐿𝑏𝑏
Equation 36

The inertia forces-

𝑠𝑠
𝑎𝑎𝐵𝐵 = 𝑥𝑥̈ ≅ − 𝜔𝜔2 (cos 𝜃𝜃 + 𝜀𝜀 cos 2𝜃𝜃)
2
Equation 37

28
In order to balance the forces of the internal components, the equation above is used with both
a positive and negative component. As shown below –

𝑠𝑠 𝑠𝑠
𝑚𝑚𝐴𝐴 �− 𝜔𝜔2 (cos 𝜃𝜃 + 𝜀𝜀 cos 2𝜃𝜃)� = −𝑚𝑚𝐵𝐵 � 𝜔𝜔2 (cos 𝜃𝜃 + 𝜀𝜀 cos 2𝜃𝜃)�
2 2
Equation 38

Acceleration of Mass A –
𝑠𝑠 2
𝜔𝜔
2
Equation 39

Inertia force resulting from Mass A –

𝑠𝑠
𝑚𝑚𝐴𝐴 𝜔𝜔2
2
Equation 40

This will have components in both the X and Y direction

𝑠𝑠 𝑠𝑠
𝑚𝑚𝐴𝐴 𝜔𝜔2 sin 𝜃𝜃 𝒂𝒂𝒂𝒂𝒂𝒂 𝑚𝑚𝐴𝐴 𝜔𝜔2 cos 𝜃𝜃
2 2
Equation 41

Therefore Fi, the total inertia force, can be either expressed in terms of its X component or its
Y component [38, 39,41]–

𝑠𝑠
𝐹𝐹𝑖𝑖 𝑥𝑥 = 𝑚𝑚𝐴𝐴 � 𝜔𝜔2 sin 𝜃𝜃�,
2
Equation 42
𝑠𝑠 𝑠𝑠
𝐹𝐹𝑖𝑖 𝑦𝑦 = 𝑚𝑚𝐴𝐴 � 𝜔𝜔2 sin 𝜃𝜃� + 𝑚𝑚𝐵𝐵 � 𝜔𝜔2 (cos 𝜃𝜃 + 𝜀𝜀 cos 𝜃𝜃)�
2 2
Equation 43

The X and the Y components of Fi also act on the block and can cause the block of the engine
to vibrate. The forces that can cause these vibrations are in equations [38, 39, 41] –

𝑠𝑠
𝐹𝐹𝑠𝑠 𝑥𝑥 = 𝑚𝑚𝐴𝐴 � 𝜔𝜔2 sin 𝜃𝜃�.
2
Equation 44
𝑠𝑠 𝑠𝑠
𝐹𝐹𝑠𝑠 𝑦𝑦 = 𝑚𝑚𝐴𝐴 � 𝜔𝜔2 sin 𝜃𝜃� + 𝑚𝑚𝐵𝐵 � 𝜔𝜔2 (cos 𝜃𝜃 + 𝜀𝜀 cos 𝜃𝜃)�
2 2
Equation 45

The V-Twin, examined in this project, had the majority of its parts modelled and therefore,
using the equations discussed in this section, allowed the con-rod to be split-up and analysed
in the methodology section.

2.3.2. Forces within the engine

When the acceleration has been calculated then the forces due to the reciprocating masses can
also be calculated. The Inertia forces are given by the equation [1, 39, 42] –

𝑠𝑠
𝐹𝐹𝑟𝑟 = 𝑚𝑚𝑟𝑟 𝜔𝜔2 (cos 𝜃𝜃 + 𝜀𝜀 cos 2𝜃𝜃)
2

29
 Equation 46

Once the speed and acceleration of the reciprocating masses has been assessed then the forces
due to these masses can be calculated using the equation [1, 7, 38, 39, 42, 44, 45] –

𝑠𝑠
𝐹𝐹𝑟𝑟 = 𝑚𝑚𝑟𝑟 𝜔𝜔2 (cos 𝜃𝜃 + 𝜀𝜀 cos 2𝜃𝜃)
2
Equation 46

Rotational mass counter weight calculation

Figure 8: Diagram of the rotating mass and balancing mass.

The equation used to calculate the counter weight is given by, [1, 2, 39, 41] –

𝐹𝐹𝐵𝐵 = 𝑚𝑚𝐴𝐴 (𝑟𝑟𝑟𝑟2 )

Equation 47

where FBal is the balancing force, ma is the mass to be balanced, r is the distance to the rotating
mass and ω is the rotational velocity

30
Once the Force to be balanced has been worked out then it must be balanced using the equation,
[1, 2, 39, 41] –
𝐹𝐹𝐵𝐵𝐵𝐵𝐵𝐵 = 𝑚𝑚𝑐𝑐𝑐𝑐−𝐵𝐵 (𝑅𝑅𝑅𝑅2 )
Equation 48

where FBal is the force needed to be countered, mcw-b is the mass needed to counter the rotating
mass, R is the distance of the balancing weight form the rotational axis, ω is the rotational
velocity.

Figure 9: Representation of forces acting on the individual engine internal components [33, 110, 111]

Gas force on the piston is required to start the calculations for the crank slider mechanism. The
equation used to find the gas force is [1, 39, 41] –

𝜋𝜋 2
𝐹𝐹𝑔𝑔 = 𝑑𝑑 𝑝𝑝
4
Equation 49

Once the gas force has been calculated then the forces acting on the internal components of the
engine can be determined. The force component Fbp can be calculated using trigonometry and
the angle ϕ. The equation calculating Fbp is [1, 39, 41] –

𝐹𝐹𝑏𝑏𝑏𝑏 = 𝐹𝐹𝑔𝑔 tan 𝜑𝜑

Equation 50

Once the component Fbp has been calculated then the driving torque acting on the engine can
be found by calculating the force component Frc acting on the big end bearing. Frc can be found
using the negative of force component Fpb acting at a distance X from the crankshaft rotational
axis. Thus the reaction torque due to Fpb can be calculated using the equation [1, 39, 41] –

𝑇𝑇𝑔𝑔 = −𝐹𝐹𝑝𝑝𝑝𝑝 𝑥𝑥 = 𝐹𝐹𝑏𝑏𝑏𝑏 𝑥𝑥 = 𝐹𝐹𝑔𝑔 tan 𝜑𝜑 𝑋𝑋

Equation 51

31
When the expression for X given above is substituted into the equation the driving torque
calculation is given by [1, 39, 41] –
1
𝑇𝑇𝑔𝑔 = 𝐹𝐹𝑔𝑔 tan 𝜑𝜑 �𝑎𝑎 × cos 𝜃𝜃 + (𝑟𝑟 2 − 𝑎𝑎2 × 𝑐𝑐𝑐𝑐𝑐𝑐 2 𝜃𝜃)2 �
Equation 52

To simplify the equation, ϕ can be taken out using the three expressions [1, 39, 41] –

𝑠𝑠 1 𝑠𝑠
sin 𝜃𝜃
tan 𝜑𝜑 = 2 𝑠𝑠 2
𝒂𝒂𝒂𝒂𝒂𝒂 cos 𝜑𝜑 = �1 − � sin 𝜃𝜃�� 𝑻𝑻𝑻𝑻𝑻𝑻𝑻𝑻𝑻𝑻𝑻𝑻𝑻𝑻𝑻𝑻𝑻𝑻 tan 𝜑𝜑 = 2 sin 𝜃𝜃
𝐿𝐿 cos 𝜑𝜑 2𝐿𝐿 1
𝑠𝑠 2
𝐿𝐿 �1 − �2𝐿𝐿 sin 𝜃𝜃��
Equation 53

Using binomial theorem to expand the three expressions results in [1, 39, 41] –

1 1
[1 − (𝜀𝜀 sin 𝜃𝜃)2 ]−2 = 1 + (𝜀𝜀 cos 𝜃𝜃)2
2
Equation 54

Rewriting the equation in terms of the square root in the denominator the higher order terms
can be neglected, this gives [1, 39, 41] –

𝑠𝑠 𝜀𝜀 2
tan 𝜑𝜑 ≅ sin 𝜃𝜃 �1 + sin2 𝜃𝜃�
2𝑙𝑙 2
Equation 55

When the expression for gas torque is substituted into the equation the driving torque is given
by the equation [1, 39, 41] –

𝑠𝑠 𝜀𝜀 2 𝜀𝜀 2 𝑠𝑠 𝜀𝜀
𝑇𝑇𝑔𝑔 ≅ 𝐹𝐹𝑔𝑔 sin 𝜃𝜃 �1 + sin2 𝜃𝜃� �𝐿𝐿 − 𝐿𝐿 + �cos 𝜃𝜃 + cos 2𝜃𝜃��
2𝑙𝑙 2 4 2 4
Equation 56

To simplify the equation it can be expanded, with all the terms of ε to a power greater than 1
neglected, the equation can be given by [1, 39, 41] –

𝑠𝑠
𝑇𝑇𝑔𝑔 = 𝐹𝐹𝑔𝑔 sin 𝜃𝜃 (1 + 𝜀𝜀 cos 𝜃𝜃)
2
Equation 57

32
 2.3.3. Analysis of gas forces
The total torque forces acting on the crankshaft from the reciprocating masses are the sum of
the torque forces, Ttot, which is the sum of both pressure Tg and inertia Ti. The Torque gas Tg
from the pressure within the cylinder can be assessed by the equation [38, 39] –

𝑠𝑠
𝑇𝑇𝑔𝑔 = 𝐹𝐹𝑔𝑔 sin 𝜃𝜃 (1 + 𝜀𝜀 cos 𝜃𝜃)
2
Equation 57
This is an approximation of the Fourier expansion from the equation [38, 39] –

𝑇𝑇𝑔𝑔 = 𝐹𝐹𝑏𝑏𝑏𝑏 𝑋𝑋 = 𝐹𝐹𝑔𝑔 tan 𝜑𝜑 𝑋𝑋

Equation 58

where [38, 39] –

𝜋𝜋 2
𝐹𝐹𝑔𝑔 = 𝑑𝑑 𝑝𝑝
4
Equation 59

The equation for torque inertia is given by [38, 39] –

𝑆𝑆 2 𝜀𝜀 3
𝑇𝑇𝑖𝑖 = −𝑚𝑚𝐵𝐵 × � � × (𝜔𝜔)2 × sin 𝜃𝜃 × � + cos 𝜃𝜃 + × 𝜀𝜀 × cos 2𝜃𝜃�
2 2 2
Equation 60

The torque inertia allows the amount of force required to start the internals rotating to be
calculated, and shows how the internals will react after a force has been applied to them. The
results of the torque inertia were input into the GT-power models to improve the accuracy. It’s
also important that the engine is properly balanced as an unbalanced engine will have a reduced
the life span, because the unbalanced forces will cause excessive wear on the internal
components. The extra loads caused by an unbalanced crankshaft will cause the engine to not
run as smoothly and to consume more fuel for a given engine load. It is possible also for internal
components to fail, with significant consequences for the engine.

33
 2.4. Manifold design and optimisation
Design change and tuning of intake and exhaust manifolds can improve engine performance,
using acoustics analysis, Helmholtz resonance theory and Ram effect [46]. This tuning process
needs to be associated with valve timing control strategies. Optimised valve timings with tuned
manifolds can offer ‘natural supercharging effect’ by using compressed inlet pressure. This
optimisation was be evaluated using a developed GT-power 1D-simulation model in order to
propose the future development of hardware.

It is well-established that a naturally aspirated engine can be tuned to have volumetric

efficiencies of over 100% using the ram-air effect.

The rules governing the whole intake within the RS rule book are [47] –

IC1.4- Air Intake System

IC1.4.1- Air Intake System Location

All parts of the engine air and fuel control systems (including the throttle or carburettor,
and the complete air intake system, including the air cleaner and any air boxes) must lie
within the surface defined by the top of the roll bar and the outside edge of the four tyres.
(See below).

Figure 10: The space where an intake is permitted to be located according to the SAE FS rules [47]

34
 2.4.1. Intake system
Optimisation of engine performance using intake and exhaust systems to increase power has
been done almost as long as internal combustion engines have been around. There are multiple
methods to calculate the optimal geometry required for a specific application. Most methods
have positives and negatives. Acoustic analysis, via Helmholtz resonance theory and the Ram
effect, can be used to improve the design of the intake and exhaust manifold. This tuning
process needs to be associated with valve timing control strategies. Optimised valve timings
with tuned manifolds can offer ‘natural supercharging effect’ by using compressed inlet
pressure, [46, 48, 49].

The simplest way of describing the acoustic waves is to apply organ tuning theory to them
using the equation below [48, 49] –

𝑐𝑐
𝑓𝑓𝑝𝑝 =
4𝐿𝐿
Equation 61

where L is the effective pipe length, c is the speed of sound and fp is the frequency of the pipe.
One of the original pieces of work in the area was done by Helmholtz [50]. The theory was
first introduced in Herman von Helmholtz’s’ book in 1862 [50]. The Helmholtz resonator
theory is an effective way of tuning intake and exhaust manifolds. The origins of the theory
come from Newton’s 2nd law, shown in Figure 11: Comparison of the similarities between a
Helmholtz resonator and a vibration absorber [49, 50].

Figure 11: Comparison of the similarities between a Helmholtz resonator and a vibration absorber

𝑑𝑑 2 𝑥𝑥
𝑚𝑚 = (𝑝𝑝𝑖𝑖 − 𝑝𝑝𝑣𝑣 )𝐴𝐴
𝑑𝑑𝑡𝑡 2
Equation 62

where m is the total mass in the runner, A is the runner cross sectional area, pi is the pressure
at the inlet, pv is pressure at the valve and x is the displacement of the air. Rearranging gives
the equation [49, 50] –

35
 𝑑𝑑2 𝑥𝑥
𝑚𝑚 + 𝑝𝑝𝑣𝑣 𝐴𝐴 = 𝐴𝐴 × 𝑝𝑝𝑖𝑖
𝑑𝑑𝑡𝑡 2
Equation 63

The assumption can then be made that the gas is acting as an ideal gas. Therefore, the adiabatic
compression can be represented as [49, 50] –

𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑
𝑑𝑑𝑑𝑑 = −𝛾𝛾𝛾𝛾 = −𝛾𝛾𝛾𝛾
𝑣𝑣 𝑉𝑉
Equation 64

The instantaneous pressure can be described by [49, 50] –

𝜌𝜌𝑐𝑐 2 𝐴𝐴
𝑝𝑝𝑣𝑣 = 𝑥𝑥
𝑉𝑉
Equation 65

The equation can then be substituted into the original expression to give [49, 50] –

𝑑𝑑 2 𝑥𝑥 𝜌𝜌𝑐𝑐 2 𝐴𝐴
𝑚𝑚 2 + 𝑥𝑥 = 𝐴𝐴𝑝𝑝𝑖𝑖
𝑑𝑑𝑡𝑡 𝑉𝑉
Equation 66

The resonator acts like a spring damper. Therefore, the equation of motion for a vibrating
system can be used, which is [49, 50] –

𝑑𝑑2 𝑥𝑥
𝐹𝐹 = 𝑚𝑚 + 𝑘𝑘𝑘𝑘
𝑑𝑑𝑡𝑡 2
Equation 67

And thus [49, 50] –

𝜌𝜌𝑐𝑐 2 𝐴𝐴
𝑘𝑘 = ,
𝑉𝑉
Equation 68

𝐹𝐹 = 𝐴𝐴𝑝𝑝𝑖𝑖 ,
Equation 69

𝑎𝑎𝑎𝑎𝑎𝑎 𝑚𝑚 = 𝜌𝜌𝜌𝜌𝜌𝜌
Equation 70

When the equations are combined, the expression below is formed [49, 50] –

𝜔𝜔 𝑐𝑐 𝑘𝑘 𝑐𝑐 𝐴𝐴
𝑓𝑓 = = � = �
2𝜋𝜋 2𝜋𝜋 𝑚𝑚 2𝜋𝜋 𝐿𝐿 × 𝑉𝑉
Equation 71

Where f is the frequency of the air, ω is the angular velocity of the engine, c is the speed of
sound, k is the harmonic order, A is the area of the throat, L is the length and V is the volume
for a given crank angle.

36
 2.4.2. Restrictor design
The rules require the use of a restrictor to limit the amount of air that can enter the engine, and
therefore the amount of power the engine can produce. The rules also limit the total
displacement. The rules that govern the restrictor are, [47] –

IC1.1.3- If more than one engine is used, the total displacement cannot exceed 610 cc and
the air for all engines must pass through a single air intake restrictor (see IC1.6 “Intake
System Restrictor.”)

IC1.6.1- In order to limit the power capability from the engine, a single circular restrictor
must be placed in the intake system and all engine airflow must pass through the
restrictor. The only allowed sequences of components are the following:
a. For naturally aspirated engines, the sequence must be (see Fig 1): throttle body,
restrictor, and engine.

IC1.6.2- The maximum restrictor diameters which must be respected at all times during
the competition are:
a. Gasoline fuelled cars - 20.0 mm (0.7874 inch)
b. E-85 fuelled cars – 19.0 mm (0.7480 inch)

IC1.6.3- The restrictor must be located to facilitate measurement during the inspection
process.

IC1.6.4- The circular restricting cross section may NOT be movable or flexible in any
way, e.g. the restrictor may not be part of the movable portion of a barrel throttle body.

IC1.6.5- If more than one engine is used, the intake air for all engines must pass through
the one restrictor.

The restrictor is a critical part of the inlet system as it governs the amount of air that can enter
the engine. It is critical in engine design to maximise the mass flowrate and minimise any losses
through the restrictor. To start assessing the inlet system a mass flowrate must be calculated
through the restrictor, as this is the start of the system and sets the upper limit of the amount of
air than can enter the inlet. [68]
𝛾𝛾+1
−
𝐴𝐴𝑡𝑡ℎ𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 × 𝑃𝑃 𝛾𝛾 𝛾𝛾 − 1 2
2(𝛾𝛾−1)
𝑚𝑚̇𝑎𝑎𝑎𝑎𝑎𝑎 = × � 𝑀𝑀 �1 + × 𝑀𝑀 �
√𝑇𝑇 𝑅𝑅 2
Equation 72

Mair is the mass flow rate of air, Throat is the cross-sectional area of throat, P is the pressure, T
is the temperature, R is the ideal gas constant, ϒ specific heat ratio and M is the Mach number.

37
 2.4.3. Runner design
The rules governing the inlet runners are [47] –

IC1.7.5 The maximum allowable Inlet Diameter (ID) of the intake runner system
between the restrictor and throttle body is 60 mm diameter, or the equivalent area (i.e.
2827 mm^2) if non-circular.

To model the pressure waves within the intake system the runners can be analysed using organ
pipe tuning theory with the pressure waves within the runner described by the equation below
[48, 49] –

𝑐𝑐
𝑓𝑓𝑝𝑝 =
4𝐿𝐿
Equation 73

where L is the effective pipe length, c is the speed of sound and fp is the frequency of the pipe.
The organ pipe tuning theory takes into account how quickly sound can travel down a pipe of
a set length but organ tuning theory does not take into account the diameter of the pipe, which
can have a significant effect on the frequency of the waves. When the pipe diameters and
volumes are taken into account the cylinder and intake can be thought of as a Helmholtz
resonator, which can be described by the equation [48, 49] –

𝑐𝑐 𝐴𝐴
𝑓𝑓𝐻𝐻 = �
2𝜋𝜋 𝐿𝐿 × 𝑉𝑉
Equation 74

where A is the cross-section of the pipe area, L is the effective pipe length, V is the volume of
the resonator, c is the speed of sound and fH is the frequency. To minimise the assumptions in
the Helmholtz resonator equation, the parameters are broken down further in the rest of this
section [49, 68] –

The effective volume of the engine is half the displacement volume as this is when the piston
is at its maximum velocity. The equation for describing the effective volume is [69] –

𝑉𝑉𝑑𝑑
𝑉𝑉𝑒𝑒 = + 𝑉𝑉𝑐𝑐
2
Equation 75

where Vd is the displacement volume and Vc is the clearance volume.

The length of the runner is ascertained from the equation, and as it is not the full length
calculated as the length of runner needs to be separated from the port length. The length
calculated is the full length on which the frequency acts, so it needs to be halved for the runner
as the wave travels to the end and back. The equation to describe the runner length is [69] –

𝑙𝑙 + 𝑙𝑙𝑝𝑝
𝐿𝐿 =
2
Equation 76

where l is the pipe/runner length and lp is the port length.

38
The velocity of the wave does not always travel at the speed of sound. It travels at the wave
propagation velocity, which can be described by [49, 69] –

𝐶𝐶𝑤𝑤 = 𝑎𝑎0 + 𝐶𝐶𝑠𝑠

Equation 77

where Cs is the practical velocity and a0 is the local speed of sound, and can be described by
the equation [49, 69] –

𝑎𝑎0 = �𝛾𝛾 × 𝑅𝑅 × 𝑇𝑇
Equation 78

where R is the ideal gas constant, γ is the specific heat ratio and T is the temperature.
The practical velocity Cs can be described by the equation [70, 71] –

𝛾𝛾−1
2 𝑃𝑃𝑖𝑖 2×𝛾𝛾
𝐶𝐶𝑠𝑠 = 2 × × 𝑎𝑎0 × � �
𝛾𝛾 − 1 𝑃𝑃0
Equation 79

where γ is the specific heat ratio, Pi is the incident pressure and P0 is the reference pressure.
Once all these equations are brought together they can be rearranged to give length –

𝐴𝐴
𝐿𝐿 = 2
𝑓𝑓
𝑉𝑉𝑒𝑒 × � 𝐶𝐶𝐻𝐻 �
𝑊𝑊
2 × 𝜋𝜋
Equation 80

The length required for the runners reduces as engine speed increases which is described by
the final equation settled on to determine the length of the inlet runners was [70, 71] –

955 × 𝑐𝑐 2 𝐴𝐴
𝐿𝐿 = � � ×� �
𝐾𝐾 × 𝑁𝑁 𝑉𝑉𝑒𝑒𝑒𝑒𝑒𝑒
Equation 81

where L is the length in mm, c is the speed of sound in m/s, K is the number of revolutions per
power cycle (2 for a four-stroke engine), N is the engine speed in RPM, A is the runner area,
and Veff is the effective volume defined by half the swept volume added to the clearance
volume. This equation is also utilised to ascertain the exhaust length.

39
 Figure 12: Comparison of wave forms of harmonic orders in a set pipe length

The harmonic orders are related to the peaks and troughs associated with the pressure waves
within a pipe. As can be seen in Figure 12: Comparison of wave forms of harmonic orders, the
waves within a pipe has have peaks, shown by B, and troughs, shown by A, which match up to
either give a pressure increase or decrease for a particular frequency at with peak or trough.
The wave length defines where these peaks and troughs are along the length of pipe and can
mean two pipes of different lengths can still give the same benefits in terms of tuning. To
ascertain which harmonic order will be most suited to the V-twin application, the equation
below was utilised, [70, 71] –

𝐿𝐿
𝐿𝐿𝑛𝑛 =
𝑛𝑛
Equation 82

where Ln harmonic length, L is the original length and n is the harmonic order.

40
 2.4.4. Plenum design
The 2018 SAE Formula Student rules that govern the plenum are [47] –

IC1.7.4- Plenums anywhere upstream of the throttle body are prohibited. For the purpose
of definition, a “plenum” is any tank or volume that is a significant enlargement of the
normal intake runner system. Teams are encouraged to submit their designs to the Rules
Committee for review prior to competition if the legality of their proposed system is in
doubt.

In an ideal plenum the pressure will be close to atmospheric (or as low as possible) to create a
pneumatic damper which will stop the pressure waves from the runner and restrictor from
interfering with each other. Minimising pressure wave interference can maximise the RAM air
effect, allowing an engine theoretically to have volumetric efficiencies of over 100%.

To assess the flow within the system the choked flow limit of the inlet port had to be assessed
using the equation [68] –
𝛾𝛾+1
𝐴𝐴𝑡𝑡ℎ𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 × 𝑃𝑃 𝛾𝛾 𝛾𝛾 + 1 −2(𝛾𝛾−1)
𝑚𝑚̇𝑎𝑎𝑎𝑎𝑎𝑎 = ×� ×� �
√𝑇𝑇 𝑅𝑅 2
Equation 72

41
 2.5. Engine simulation literature review

1D engine simulation is an essential step in engine design and development. 1D simulations

are essentially simplified 3D models and only deal in a single plane or direction. 1D simulation
is commonly used to evaluate parameters such as valve dimensions and lift, as well as manifold
design, [73]. If there is no access to experimental combustion data [81] has stated that a 1-D
non-predictive model is the most accurate way to simulate the combustion event. [81] goes on
to say that GT-Power was the chosen software for its versatility and extensive library of internal
combustion engine components.

1D engine simulation commonly requires the building of a models composed of sub-models,

each running a specific task/calculation. During this ‘Block’ building assumptions or
simplifications must be made for each sub-model before the simulation can be run. 1D engine
simulation is not a ‘one stop shop’ but the first step in the design process. Whilst 3D simulation
might give more accurate results, the significant advantages in using 1D are that it is faster and
less demanding of computing power, and therefore also less expensive. 1D engine simulation
software allows for the quick identification of potential design problems and can highlight areas
that need improving without the time and cost involved in developing a full 3D model, [73].

It was found by [75] that a non-predictive model in GT-Power could cut simulation times by
more than 50% compared to a predictive model. One drawback of 1D simulation is the limited
solvers unless implemented by external software such as Matlab or Simulink. However, engine
development is an iterative process for which 1D analysis has been found to be an excellent
starting point, [72, 73]. One limitation found by [76] was that the MFB profile calculated by
the single Wiebe model cannot account for a combustion that has different rates of combustion
before and after 50% MFB. This limitation can be calibrated by a double Wiebe function model
within GT-Power, although real-time data is needed to calibrate the double Wiebe model while
the single Wiebe model does not need the real-time data. [77] went onto say that although using
the double Wiebe model increased the accuracy of the model is also increased computational
complexity. When calculating a predictive model, engine modelling software was found by
[76] to do a poor job of capturing transients and discontinuities, which is more evident when
calculating stochastic engine operation. Whereas the effects of parameters such as BSFC, CR,
fuel composition, BMEP, Knock and volumetric efficiency can be shown effectively by the
simulations. [77] states that, due to the Wiebe model computing parameters as a function of
engine geometries and operating conditions through physically based parameters, the user can
have confidence in the simulations results for GT-Power, while using any of the heat release
method of combustion simulation within GT-Power requires accurate in-cylinder temperature
data for comparison and calibration.

There are multiple 1D engine simulation software products available on the market. Well
known examples include GT-Power, Ricardo Wave and Virtual Engines. The software chosen
for this project is GT-Power, which is a part of the GT-Suite simulation family of software,
made up of multiple software packages which analyse different parts of the engine and
powertrain. Each of the packages within GT-Suite is based upon a common library of
equations, leading to the benefit that the multiple stages of engine development simulation can
be integrated together into a coherent package. Simulations run on GT-Suite can also be
integrated with simulations run on other software products, such as STAR-CD, Simulink,
Matlab and EXCEL, [74].

42
GT-Power is, when calibrated correctly, known to be reliable based on evidence from a large
user population who have tested and validated its real-time accuracy, [92, 93, 94]. Meyer, J,
[79], states that engine simulation software should be used alongside engine test cell
calibration, as a well calibrated model can allow more in-depth analysis of the engine than
engine calibration in a test cell alone. This is due to GT-Power being able to calculate the
residual gas fractions which would otherwise be very hard to measure accurately. Although
GT-Power recommends between 25-200 points of data to calibrate the engine models while
Kumar, M, et al, [80], only used 25 points due to time constraint. The project had an accuracy
limit for the models of within ±5% efficiency accuracy and managed average IMEP inaccuracy
of ±2.2% when compared to measured in-cylinder data. Hatlevold, E, S, [78], found that, with
calibration from in-cylinder data, ±2% accuracy could be achieved with a GT-Power simulation
model. Although Hatlevold goes onto say that it was common to find a discrepancy of ±4%
when comparing peak pressure with real-time data.

In order to start the development of a GT-Power model, there are parameters that need to be
known in order to start the process. The parameters that were needed to start the construction
of the model can be seen in Table 2: Initial Engine Specs for Model Development, [82] –

section parameter value unite reference

engine characteristics compression ratio 12.45 ratio measured from engine
firing order 1 then 2 - measured from engine
engine configuration V-Twin - measured from engine
angle of cylinders 75 deg measured from engine
2 or 4 stroke 4 stroke - measured from engine
cylinder geometry bore 92 mm measured from engine
stroke 45.13 mm measured from engine
con-rod length 103 mm measured from engine
pin offset 0 mm measured from engine
piston bowl geometry - - input from CAD models
piston area - - input from CAD models
head area - - input from CAD models
intake and exhaust system geometry of manifolds - - input from CAD models
runners 311 mm measured from engine
exhaust 614 mm measured from engine
ports - - input from CAD models
throttles location start of intake - input from CAD models
fuel injectors number 2 - measured from engine
location Port injection - measured from engine
discharge coefficient - - given by manufacturer
air - fuel ratio 14.7 ratio stoichiometric
fuel type pump gasoline - limited by SAE FS rules
intake and exhaust valves Intake valve lift 10.978 mm input from CAD models
exhaust valve lift 9.727 mm input from CAD models
lift profile - - input from CAD models
discharge coefficients - - calculated parameter
ambient state pressure 1 bar Ambient conditions
temperature 300 K Ambient conditions
Table 2: Initial Engine Specs for Model Development

Some of the critical outputs that can be extracted from an engine simulation are Power, torque,
IMEP, BSFC and Volumetric Efficiency.

43
 2.6. V-twin engine performance and calibration
2.6.1. Knock
Knocking or detonation is one of the main factors limiting modern SI engine performance.
Knock is where the air/fuel charge within the cylinder self-combusts, causing higher
temperatures and pressures in the cylinder. This can have a degrading effect on engine
performance and can cause extreme pressure on components, and their wear and even failure.
Knock detection is a critical part of engine development and engine performance optimisation.
The leaner an engine runs the more likely that knock will occur due to the increased likelihood
that pre-ignition or Compression Auto Ignition (CAI) would happen within the cylinder. Pre-
ignition or CAI causes a sharp increase in pressure in the cylinder before the piston is at TDC.
The increase in pressure can cause excessive pressure in the cylinder and can result in severe
engine component damage. In effect, maximum fuel efficiency is achieved in conditions in
which an engine is more likely to knock, as both are a function of less fuel being injected per
cycle.

Ideally combustion starts at the sparkplug and propagates outwards towards the cylinder wall
until the fuel/air charge is fully burnt, giving smooth and controlled combustion. The rate at
which the charge burns gives the engine one of its performance characteristics. The optimal
crank angle, after TDC, for peak pressure, and the time taken for the charge to burn, give an
indication of which crank angle, before TDC, the fuel should be ignited. If the charge ignites
before that given crank angle, or the fuel combusts quicker than its calculated time, the fuel
can ignite as the piston is still compressing the air/charge inside the cylinder. This is knocking
and can cause a peak in pressure and temperature far higher than it would otherwise have been.
In some cases, the pressure inside the combustion chamber can be high enough to damage the
internal components.

The higher the performance of the engine, the more likely there will be the conditions for
knocking to occur. To reduce the risk of knock, parameters such as spark timing, injection
timing, injection duration, CAM timing and AFR can be adjusted. Alternately these parameters
can be kept at levels more consistent with optimum performance and knocking controlled with
an anti-knock strategy.

Knock detection

There are two main ways to detect knock. The first is analysis of the in-cylinder pressure trace
and the other is vibration or noise analysis of the engine. Within a non-knocking engine cycle,
the heat released from the fuel follows a predictable profile. Within a knocking engine cycle
the heat release rate has a measurably different profile. If knock is detected then anti-knock
strategies can be put in place to limit it.

Vibrational or noise knock detection

Vibrational knock detection is usually done using a vibration sensor attached on the outside of
the block. Vibrational analysis is a cheaper way of detecting knock, but it is not as accurate as
the analysis of an in-cylinder pressure trace.

44
In-cylinder pressure knock detection

In-cylinder pressure knock detection is done using an in-cylinder pressure trace, which is an
expensive but more accurate way of detecting knock.

Knock detection
One method of detecting auto ignition and knocking conditions is comparing the oscillations
of the in-cylinder pressure trace, and is done using the equation [30, 31, 32] –
0
𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 = max[|𝑝𝑝̂ |]𝑇𝑇𝑇𝑇𝑇𝑇+𝑁𝑁
𝑇𝑇𝑇𝑇𝑇𝑇
Equation 83

where MAPO is the Maximum Amplitude of Pressure Oscillations, 𝑝𝑝 � is the filtered in-cylinder
pressure/pressure osculation signal and N is the number of crank angles chosen to compare.

Figure 13: A comparison of Maximum amplitude of pressure oscillations over knocking cycles, [12].

Another method is using the HRR and comparing it to MFB50. When the auto ignition happens
before MFB50 then the fuel/air mixture combusting in front of the flame front causes rapid
increases in pressure and temperature, which can be detected using an in-cylinder pressure
trace. Auto-ignition will also shorten the combustion duration and change the crank angle at
which MFB50 takes place. The equation used for the HRR is [30, 31, 32] –

𝑑𝑑𝑑𝑑 1 𝑑𝑑𝑑𝑑 𝛾𝛾 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑ℎ𝑡𝑡

= 𝑉𝑉 + 𝑝𝑝 +
𝑑𝑑𝑑𝑑 𝛾𝛾 − 1 𝑑𝑑𝑑𝑑 𝛾𝛾 − 1 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑
Equation 14
Q is the combustion heat energy release, Qht is the heat energy transfer to the chamber wall, V
is the volume, θ is the crank angle and γ is the specific heat ratio. The time of auto-ignition is
estimated by the solution of taA is given by the solution of [30, 31, 32] -
𝑡𝑡𝑎𝑎𝑎𝑎
𝑑𝑑𝑑𝑑
1=�
𝑡𝑡𝑆𝑆𝑆𝑆𝑆𝑆 𝜏𝜏𝑐𝑐ℎ
Equation 84

45
where taA < tEOC and tSOC is the start of combustion time and tEOC is the end of combustion time.
For Hydrocarbons τch is expressed by [30, 31, 32] –

𝐶𝐶𝑎𝑎3
𝜏𝜏𝑐𝑐ℎ = 𝐶𝐶𝑎𝑎1 × 𝑝𝑝−𝐶𝐶𝐶𝐶2 × 𝑒𝑒𝑒𝑒𝑒𝑒 � �
𝑇𝑇𝑒𝑒
Equation 835

where Te is the end-gas temperature. Ca1, Ca2 and Ca3 are fuel Octane Number which are
given as Ca1 = 17.68x10-3, Ca2 = 1.7 and Ca3 is 3800 according to [32] showing the higher the
octane of the fuel used the more resistant it is to knock in the first place.

Figure 14: A comparison of knocking and non-knocking pressure traces, used to identify knocking engine
cycles [1].

Another method of analysing knock is to detect the resonance, caused by auto-ignition. This
resonance, which can occur after TDC and towards the end of combustion, can be detected
within the cylinder, and can be calculated from in-cylinder pressure using the equation, [30,
31, 32] –

𝜕𝜕 2 𝑝𝑝
= 𝐶𝐶 2 ∇2 𝑝𝑝
𝜕𝜕𝑡𝑡 2
Equation 846

where t is time, C is the speed of sound and ∇2 𝑝𝑝 is the Laplacian equation regarding change in
pressure p, [30, 31, 32].

Anti-knock strategy

One of the easiest and quickest ways to reduce knock, is to retard the spark timing. This method
can be applied individually to each cylinder, allowing performance to be optimised for each
cylinder. This anti-knock strategy doesn’t have to limit performance in all the cylinders to the
same degree, therefore allowing a better overall performance.

46
Another method of countering knock is to delay the fuel injection timing as this will have a
similar effect to the retarding the spark timing, but it can only be done on an engine with Direct
Injection. Port Injection is reliant on the valve timing.

Another method is to use a higher Octane of fuel as higher Octane fuels have improved anti
knock properties. Pump fuel ranges from 95 Octane to 101 Octane, and in this study the highest-
Octane pump fuel of 101 Octane was be used as it helps reduce knock as well as improve
performance.

Water injection can also allow in-cylinder temperatures to be reduced and therefore is another
method to prevent engine knock.

Integration into the GT-Power model

All of the combustion profiles that can be selected, when constructing a GT-Power model, can
utilise a predictive knock model by selecting a knock reference object within the selected
combustion profile. The knock model reference objects allow phenomenological predictions
of knock and/or pre-ignition tendencies based on multiple empirical correlations for induction
time, based on work done by Worrent, Franzke and Eyzart et al [106, 107, 108]. However, the
quantitative predictions of knock must be calibrated with experimental engine measurements,
[10].

2.6.2. Valve Train

The valve train can be split into 6 main areas –

Valve Lift (VL) – The VL directly affects the performance of the engine, with a larger VL
giving higher performance. This has been linked in some literature to the viscous effect of the
attached jet having proportionately more impact with a smaller VL [7].

Inlet valve opening (IVO) – IVO has a large impact on engine performance as it directly affects
the amount of air that can enter the engine [2]. Retarding the IVO past the piston TDC can
help the inertia of the fluid to be overcome in the intake manifold, which should lead to an
increase in volumetric efficiency [84, 85, 86]. In Reference [84] it was found that there was an
increase in torque with IVO before Top Dead Centre (bTDC), with 20 degrees CA bTDC giving
the best results. Retarding the IVO further caused the power to fall off again, the drop in power
being assessed to be due to an increase in Exhaust Gas Recirculation (EGR), with lower
performance due to some of the charge in the cylinder being made up of exhaust gas thus giving
less oxygen to burn.

Inlet valve closing (IVC) – IVC has a large impact on engine performance and should take
account of the application for which the engine is being designed [1, 86]. The ideal point of
IVC is when there is the maximum volume of air in the cylinder. Following the ideal engine
cycle, optimum IVC is at Bottom Dead Centre (BDC), although at this point there is still air
coming into the engine from the natural supercharging effect caused by pressure differences in
the engine. Taking into account the reduced time for air to enter the cylinder at higher engine
speeds, IVC should be retarded to maximize the performance of the engine [45]. At lower
engine speeds and when the engine isn’t at wide-open throttle/fully loaded, the inlet IVC ideally
should be advanced to reduce the pumping losses within the engine [88].

47
Exhaust valve opening (EVO) – EVO is a trade-off between extracting the maximum amount
of exhaust gas from the cylinder and being able to extract the maximum amount of work from
the engine delivered by expansion of the burnt charge [89]. The knock-on effect of an increase
in exhaust gas in the cylinder at IVO is that there is a higher pressure in the cylinder and a
reduction in the natural supercharging effect and therefore lower volumetric efficiencies [86,
89]. Another factor that has to be taken into account is that at higher engine speeds there is less
time for the exhaust gas to be extracted and therefore the EVO should ideally be retarded.
Reference [85] found that taking into account the exhausting of the burnt charge, the optimal
EVO for race engines is about 90 degrees CA bBDC, and for road engines around 50-60
degrees CA bBDC. The EVO CA varies with engine speed as well as load, as both of these
factors affect the in-cylinder pressure. A reduced in-cylinder pressure benefits from a retarded
EVO, [86].

Exhaust valve closing (EVC) – EVC affects the amount of Exhaust Gas Recirculation (EGR)
that is present in the engine. Advancing the EVC can increase the amount of exhaust gases in
the cylinder at lower engine speeds, as indicated in [90]. This is done to improve the emissions
of the engine. The negative effect of advancing the EVC, however, is that it reduces the
performance and can cause instability in the combustion event, leading to knock from the
increased combustion temperatures. At higher engine speeds an advanced EVC leads to higher
volumetric efficiencies due to the natural supercharging effect, as the air is drawn in by the
lower pressures in the exhaust and EGR is not allowed to happen as the valve closes at the
optimal time [84, 85, 86].

Valve overlap (VO) – VO has to be controlled carefully as, depending on the engine speed, the
VO can either be beneficial or detrimental to engine performance. At high engine speeds the
overlap draws more air into the cylinder as, if the inlet frequencies are tuned right, there is
lower pressure in the cylinder and even lower pressure in the exhaust manifold. A larger intake
overlap benefits higher engine speeds while a smaller overlap benefits lower engine speeds. At
lower engine speeds if there is too much overlap, there is a negative effect on the pressure
differential. However, at higher engine speed, more time is needed for the effects to be seen,
so a larger overlap is needed, which can increase volumetric efficiencies. However, at low
engine speeds this effect can draw the air out into the exhaust causing lower volumetric
efficiency [2, 5, 90].

Further work will look into the potential of variable valve timing, as this can offset some of
the problems associated with having fixed valve timing. Variations in timing can potentially
deliver benefits across the rev range and allow valve timing to be optimised at multiple engine
speeds [101, 102].

48
 2.7. Computational Fluid Dynamics (CFD)

Introduction
CFD is the solving of fluid dynamic problems via numerical analysis and algorithms which
allow complicated fluid flows to be accurately predicted by computers. Computers are needed
because of the complexity of the problems. Modern CFD software can not only calculate the
fluid flow behaviour but also parameters such as heat transfer, mass and mechanical
movements.

Navier-Stokes
The majority of Computational Fluid Dynamics (CFD) packages on the market utilise the
Reynold-Averaged Navier-Stokes (RANS) equations which are the core of CFD. CFD
packages have grown in complexity since the 1980s, especially with the physical modelling
but with less emphasis on geometric complexity. On the other hand Computer Aided Design
(CAD) have become commonplace, if not crucial, to the design process within industry with
modern CAD packages allowing the development of complex shapes to be constructed with
ease. In 1990 the first Mechanical Computer Aided Design (MCAD) system had a CFD
package, SolidWorks FloWorks, incorporated into it. This allowed the CFD package to use the
native CAD geometry without any required modification at the start of the process, [92].
SolidWorks FlowSim solves the Navier-Stokes Equations in the fluid region, which are
formulations of Momentum, Mass and energy conservation Laws [91, 92, 93, 94].

The Continuity equation. A three-dimensional continuity equation for a compressible fluid can
be used to describe the conservation of mass, [98]. –

𝜕𝜕𝜕𝜕 𝜕𝜕(𝜌𝜌𝑢𝑢1 )
+ =0
𝜕𝜕𝜕𝜕 𝜕𝜕𝑥𝑥𝑖𝑖
Equation 857

where ρ is the fluid density, t is the temperature and u is the fluid velocity. The first term is the
rate of change of density over the change in time. The net mass flow of the elements across the
boundaries is described by the second term, [93, 99].

The Momentum equation. The rate of change of momentum can be described, according to
Newton’s second law, by the momentum equation known as the Navier-stokes equation, [98]
–

𝜕𝜕𝑢𝑢𝑖𝑖 𝜕𝜕𝑢𝑢𝑖𝑖 1 𝜕𝜕𝜕𝜕 1 𝜕𝜕𝜏𝜏𝑗𝑗𝑗𝑗

+ 𝑢𝑢𝑗𝑗 =− + + 𝑔𝑔𝑖𝑖
𝜕𝜕𝜕𝜕 𝜕𝜕𝑥𝑥𝑗𝑗 𝜌𝜌 𝜕𝜕𝑥𝑥𝑖𝑖 𝑝𝑝 𝜕𝜕𝑥𝑥𝑗𝑗
Equation 868

where g is the gravitational accelerator, ρ is the fluid density, τ is the Viscous stress Tensor, t
is the fluid temperature and u is the fluid velocity, [92, 98].

The Energy equation. Internal energy, U, in an incompressible fluid for a constant cp can be
describe by, [98] –

𝑑𝑑𝑑𝑑 𝜕𝜕𝑢𝑢𝑖𝑖 𝜕𝜕 𝜕𝜕𝜕𝜕

𝜌𝜌 = −𝑝𝑝 + 𝛷𝛷 + �𝑘𝑘𝑚𝑚 �
𝑑𝑑𝑑𝑑 𝜕𝜕𝑥𝑥𝑖𝑖 𝜕𝜕𝑥𝑥𝑖𝑖 𝜕𝜕𝑥𝑥𝑖𝑖
Equation 879

49
where g is the gravitational accelerator, ρ is the fluid density, τ is the Viscous stress Tensor, t
is the fluid temperature and U is the fluid velocity [92, 98]. Where –

2
𝛷𝛷 = 2𝜇𝜇𝑆𝑆𝑖𝑖𝑗𝑗 𝑆𝑆𝑖𝑖𝑖𝑖 − 𝜇𝜇𝑆𝑆𝑖𝑖𝑖𝑖 𝑆𝑆𝑘𝑘𝑘𝑘
3
Equation 90

where µ is the dynamic viscosity and Sij is the strain-rate tensor, given by, [92, 98] –

1 𝜕𝜕𝑢𝑢𝑖𝑖 𝜕𝜕𝑢𝑢𝑗𝑗
𝑆𝑆𝑖𝑖𝑖𝑖 = � + �
2 𝜕𝜕𝑥𝑥𝑗𝑗 𝜕𝜕𝑥𝑥𝑖𝑖
Equation 881

where g is the gravitational accelerator, ρ is the fluid density, τ is the Viscous stress Tensor, t
is the fluid temperature and u is the fluid velocity [92, 98].

Meshing
CAD software is used to define solid models, whereas CFD software is used to describe the
flow space. More commonly in CFD the Fluid space is created by Boolean subtraction of a
solid model within CAD. This is then passed onto the CFD in order to mesh the model, [92].
Meshing can be done via irregularly distributed nodes for complex shapes, often referred to as
unstructured mesh, or by structured mesh for less complex shapes. A combination of the two
can be used, with unstructured meshes used around complex shapes and near boundaries, and
the structured mesh used everywhere else. This is often referred to as partially structured mesh.
This process is done via a mesh generator built into most CFD software [92, 98].

Figure 15: a shows of irregularly distributed mesh, b shows structured distributed mesh and c shows
partially structured mesh, [92].

The start of the meshing process usually happens at the surfaces, by means of Delaunay
triangulation [96], then the space is meshed via the surface triangulation. The mesh is often
generated with tetrahedral elements [97] that meet the Delaunay criterion [96]. It is often the
case that, once the mesh has been generated, the user will have to solve any ambiguities and/or
defects within the mesh or sort out any over-defined areas [92, 93, 94].

50
 Figure 16: a is a first order tetrahedral element with 4 nodes, b is a degenerate first order tetrahedral
element with 8 nodes, c is a second order tetrahedral element with 10 nodes, d is a degenerate second
order tetrahedral with 20 nodes, [6].

The nodes lie on the intersection of the boundaries of the cells and are normally the starting
points for the construction of the elements [97, 100].

Boundary conditions
The boundary conditions of the CFD simulation are an important part of setting up a simulation
and are critical to the accuracy of the results. They can be split into different sections

Inlet and outlet boundaries

The inlet and outlet boundaries describe parameters such as Mass Flow Rate and velocities that
are either entering or being drawn out of the fluid body. They normally describe the start of the
simulation, but they can also be the parameter that is being measured if there is an event
happening within the fluid body, [92, 93, 96, 98, 99].

Wall boundaries
The wall boundary is the edge of the fluid body that is not an inlet or outlet. The wall conditions
describe how the fluid interacts with the wall in terms of parameters such as friction and shear
force, allowing the turbulence and boundary layers of the fluid to be calculated within the fluid
body. The heat transfer to and from the walls is also a critical parameter for accurate CFD
simulations, [92, 93, 97, 98, 99].

Symmetry and axis boundary conditions

The common way of reducing simulation time and computing power required within industry
is to halve a symmetrical model. Although this is may reduce the accuracy of a simulation, due
to time constraints of processing power of a computer, this may be a preferable way of
simulation. However as a lot of fluid flow simulations, especially turbulent or non-symmetrical
inlet conditions, can give a number of possible simulation results, this may reduce the accuracy
further, [92, 93, 95, 98, 99].

Initial conditions
The initial conditions boundaries relate to how the fluid in the fluid body is acting before the
simulation starts, as this can have an effect on the results. More accurate initial conditions can
improve the accuracy as well as the time taken for the simulation to run. Initial conditions also
relate to the ambient conditions of the fluid entering the fluid body, [92, 93, 98, 99].

51
 Chapter 3.Methodology

52
 3. Methodology
3.1. Combustion
The combustion process will be analysed using combustion phasing, as this allows the
combustion process to be broken down and analysed without going into a chemical kinetic
combustion model, which would be too complex and time consuming. Representing the in-
cylinder combustion through combustion phasing will highlight key areas of the combustion
and allow for further analysis. Combustion phasing will be represented by the use of Heat
Release Rate (HRR) and Mass Fraction Burn (MFB). Analysing combustion phasing through
HRR and MFB will allow different areas of combustion to be analysed, such as the efficiency
of the combustion and where optimal peak pressure should ideally be within the cycle.

The combustion model theory was combined into an Excel template to analyse the combustion
process and allow for further tests of the different Wiebe model parameters to be incorporated
into the GT-power model. Figure 18: A representation of the combustion model constructed in Excel.
shows a representation of the Excel model, the figures from the model have been shown in
detail throughout the project so a detailed view was not deemed necessary.

Figure 17: Otto cycle Pressure Volume profile

The PV diagram shown in Figure 17: represents the combustion stroke of the Otto cycle. The
start of the cycle is at 1, the reduction in volume and increase in pressure comes from the piston
compressing the air inside the cylinder and finishes at 2. From 2 to 3 is the heat/pressure
addition at constant volume from combustion. 3 to 4 is the reduction of pressure and increase
in volume as the piston moves back down the cylinder. 4 to 1 is back to the start of the cycle
but doesn’t include the exhaust and the intake stroke which accounts for the change in pressure
due to the intake of ambient air. When the value for energy release is calculated then the Otto
cycle trace can be calculated and the results put into an Excel Spreadsheet in order to allow
analysis of the engine and the combustion model. The Otto cycle, for the compression and

53
expansion stroke, uses the ideal gas equation expression pVγ=constant and is calculated using
the following equations [1, 7, 41, 45] –

𝑉𝑉𝑐𝑐 + 𝑉𝑉𝑑𝑑 γ 𝑉𝑉𝑐𝑐 + 𝑉𝑉𝑑𝑑 γ−1

𝑃𝑃(𝜃𝜃) = 𝑝𝑝0 � � 𝑎𝑎𝑎𝑎𝑎𝑎 𝑇𝑇(𝜃𝜃) = 𝑇𝑇0 � � 𝑓𝑓𝑓𝑓𝑓𝑓 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐
𝑉𝑉(𝜃𝜃) 𝑉𝑉(𝜃𝜃)
Equation 892

𝑉𝑉𝑐𝑐 γ 𝑉𝑉𝑐𝑐 γ−1

𝑃𝑃(𝜃𝜃) = 𝑝𝑝3 � � 𝑎𝑎𝑎𝑎𝑎𝑎 𝑇𝑇(𝜃𝜃) = 𝑇𝑇3 � � 𝑓𝑓𝑓𝑓𝑓𝑓 𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒
𝑉𝑉(𝜃𝜃) 𝑉𝑉(𝜃𝜃)
Equation 903

The Otto cycle assumes that there is ideal heat addition which means that all the heat that can
be released by the fuel/air charge is released at the optimal point. Due to the compression at
TDC, this is assumed to be the ideal point for the energy to be released, although it is a little
after TDC as this is when the piston starts to move back down the cylinder and therefore comes
under least resistance from the forces of inertia. Energy release must be worked out using the
equations [1, 7, 41, 45] –
𝑚𝑚𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 ℎ𝐿𝐿𝐿𝐿𝐿𝐿 𝑇𝑇3
𝑇𝑇3 = 𝑇𝑇2 𝒂𝒂𝒂𝒂𝒂𝒂 𝑝𝑝3 = 𝑝𝑝2
𝑚𝑚𝑎𝑎𝑎𝑎𝑎𝑎 𝑐𝑐𝑣𝑣𝑣𝑣 𝑇𝑇2
Equation 914

After TDC the Otto cycle follows isentropic expansion with the equations [1, 7, 41, 45] –

𝑚𝑚𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 ℎ𝐿𝐿𝐿𝐿𝐿𝐿 𝑇𝑇3

𝑇𝑇3 = 𝑇𝑇2 + 𝒂𝒂𝒂𝒂𝒂𝒂 𝑝𝑝3 = 𝑝𝑝2
𝑚𝑚𝑎𝑎𝑎𝑎𝑎𝑎 𝑐𝑐𝑣𝑣𝑣𝑣 𝑇𝑇2
Equation 925

The Mass Fraction Burn by McCuiston, Lavoie and Kauffman (MLK) model, [109], can be
used to create a experimental combustion model or a theoretical combustion model can be
constructed using the Wiebe method. An experimental version matches known data, such as a
pressure trace, in order to construct the model. Whereas a theoretical model uses the
fundamental equations in order to construct the model. Once the Qcycle has been worked out,
then the heat release can be calculated using the Wiebe function over the combustion phase [1,
7, 41, 45] –

𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑎𝑎𝑎𝑎𝑎𝑎
𝑄𝑄𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶 = 𝑚𝑚𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 ℎ𝐿𝐿𝐿𝐿𝐿𝐿 = ℎ𝐿𝐿𝐿𝐿𝐿𝐿
𝐴𝐴𝐴𝐴𝐴𝐴
Equation 936

Once the Qcycle has been worked out, then the heat release can be calculated, using the Wiebe
function, over the combustion phase [1, 7, 41, 45] –

𝑄𝑄𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 𝜃𝜃 − 𝜃𝜃0 𝑚𝑚+1

𝑥𝑥𝑏𝑏 = = 1 − exp[ −𝑎𝑎 × � � ]
𝑄𝑄𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 𝛥𝛥𝛥𝛥
Equation 19

Where xb is the mass fraction burn, θ0 is the crank angle at the start of combustion, and Δθ is
the crank angle duration of combustion. The typically for a and m for an SI engine are a=5 and
m=3. In order to show the dQ/dθ in terms of J/Rad [1, 7, 41, 45] –

54
 𝑑𝑑𝑑𝑑 𝑄𝑄𝑛𝑛 − 𝑄𝑄𝑛𝑛+1
=
dθ θ𝑛𝑛 − θ𝑛𝑛+1
Equation 947

The parameters m, a and b were investigated and compared to those for similar engines to
create a more accurate combustion model and improve the simulations run in GT-power, [19,
20].

For the shaft work of the engine, [1, 7, 41, 45] –

𝑑𝑑𝑑𝑑𝑠𝑠ℎ𝑎𝑎𝑎𝑎𝑎𝑎 (𝑝𝑝 − 𝑝𝑝0 )𝑑𝑑𝑑𝑑

𝑊𝑊̇𝑠𝑠ℎ𝑎𝑎𝑎𝑎𝑎𝑎 = = = 𝜏𝜏𝜏𝜏
𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑
Equation 958

Where τ is the toque and ω is the angular velocity, [1, 7, 41, 45].

In order to get the torque independent of velocity then divide through by angular velocity ω,
and because ω=θ/t, to get [1, 7, 41, 45]-

𝑑𝑑𝑑𝑑𝑠𝑠ℎ𝑎𝑎𝑎𝑎𝑎𝑎 (𝑝𝑝 − 𝑝𝑝0 )𝑑𝑑𝑑𝑑

𝜏𝜏 = =
𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑
Equation 969

Figure 18: A representation of the combustion model constructed in Excel.

55
The combustion model allowed critical factors of the combustion event to be analysed post
simulation and allowed for a greater understanding of the combustion event, the outcomes of
the model development were –

Figure 19: Volume displacement profile from Excel combustion model.

The starting point of the of the combustion model was to get a volume displacement profile, as
this allowed the pressure change inside the cylinder, due to compression, to be calculated.

Figure 20: Pressure comparison of the Otto cycle and Wiebe function.

The next step in the combustion model development was to calculate the pressure change due
to both the compression of the air from the piston and the combustion of the fuel.

56
 Figure 21: Temperature comparison of Otto cycle and Wiebe functions.

From the pressure change, the temperature change could be calculated

Figure 22: Pressure Vs Volume graphs for Wiebe function and Otto Cycle.

Once the temperature and pressure change had been calculated, then the pressure vs volume
diagram could be constructed. This is a very useful tool in analysing the combustion event.

57
 Figure 23: Shaft work for both Otto cycle and Wiebe function.

The shaft work gives an indication of how much positive and negative work is being done by
the combustion event onto the piston and can give power performance indications.

Figure 24: Torque per crank angle for the Otto cycle and Wiebe function.

The Torque per CA graph is directly related to the shaft work graph and is related and shows
the instantaneous torque applied to the piston per crank angle.

58
 Figure 25: dQ/dθ profile.

The dQ/dθ profile shows how much energy is realised for the fuel per crank angle and is directly
linked to the MFB curve. Although most simulations calculate the dQ/dθ from the MFB curve
it is still a critical factor in understanding the combustion event.

Figure 26: MFB profile for the Wiebe function.

The mass faction burn profile is the critical outcome of the combustion model. It allows the
combustion event to be compared and contrasted over different combustion events and shows
the potential power performance of an engine.

59
 3.2. Dynamics
The engine studied in this project is based on 2 cylinders of a Cosworth DFV Formula 1 V8.
The engine dynamics had to be checked in order to be sure that parameters such as the con-rod
length, crankshaft counterweight and crank offset where validated for the engine. The CAD
software CATIA R5V24 was used to location the Centres of Mass for the current components
among other parameters.

One of the key parameters examined was Mean Piston Speed (MPS), which is a key area in
engine development as the higher the MPS, the quicker the flame propagation happens within
the cylinder. This can be seen in the flame propagation research section of the Appendix.
However, the higher the MPS the larger the internal frictions and therefore the larger the
mechanical efficiency losses and fatigue and wear of components. However, MPS also affects
volumetric flow rate of air into the cylinder, as at lower speeds more air can be transferred into
the cylinder per cycle. However, this is countered by a lower pressure differential between
outside and inside the cylinder which acts to reduce airflow into the cylinder.

Piston acceleration is also an important factor in analysing the forces acting on the internal
components such as the piston or journal bearings. A dynamics model constructed in an Excel
spreadsheet was used to ascertain the forces acting on the crank slider mechanism using the
equations below. Reducing the peak acceleration of the piston will result in reduced forces
acting on the engine internals. This is because F=ma and, as the mass remains constant,
reducing the acceleration will also reduce the forces, thus reducing wear and fatigue. Reduced
acceleration of the piston will also result in more time for gas exchange.

Taking all of these factors into account the selected MPS will be a compromise between the
benefits and disadvantages of high and low piston speeds. For this project an MPS of 30m/s
was chosen which is consistent with the previous research and development of this engine, [1,
26, 27, 39, 41, 42].

The data collected from the CAD models was input into the Excel spreadsheets using the
equations from the literature review. Figure 27: A representation of the dynamics model constructed
in Excel. shows a representation of the Excel model. The figures from the model have been
shown in detail throughout the project so a detailed view is not deemed necessary. The
following equations where used to construct the Excel model -

Piston displacement –
1
𝑠𝑠 𝑠𝑠 2 2
𝑋𝑋 = cos 𝜃𝜃 + 𝐿𝐿 �1 − � sin 𝜃𝜃� �
2 2𝐿𝐿
Equation 28
Piston velocity [1, 39, 41, 42] –

𝑠𝑠 𝜀𝜀
𝑋𝑋̇ = − 𝜔𝜔 �sin 𝜃𝜃 + sin 2𝜃𝜃�
2 2
Equation 29
Piston acceleration [1, 39, 41, 42] –
𝑠𝑠
𝑋𝑋̈ = − 𝜔𝜔2 (cos 𝜃𝜃 + 𝜀𝜀 cos 2𝜃𝜃)
2
Equation 35
where X ̈ is the approximation of the piston acceleration,

60
Gas Torque [38, 39] –

𝑠𝑠
𝑇𝑇𝑔𝑔 = 𝐹𝐹𝑔𝑔 sin 𝜃𝜃 (1 + 𝜀𝜀 cos 𝜃𝜃)
2
Equation 57
Torque inertia [38, 39] –

𝑆𝑆 2 𝜀𝜀 3
𝑇𝑇𝑖𝑖 = −𝑚𝑚𝐵𝐵 × � � × (𝜔𝜔)2 × sin 𝜃𝜃 × � + cos 𝜃𝜃 + × 𝜀𝜀 × cos 2𝜃𝜃�
2 2 2
Equation 60

The project will cover analysis of the forces over 4k,9k,and 12k RPM in order to analyse the
forces but particular attention will be paid to 9k RPM as this is used to located peak torque.

Figure 27: A representation of the dynamics model constructed in Excel.

Figure 28: Piston displacement.

The start of the dynamics model is to calculate the distance the piston is down the stroke. This
will allow other parameters to be calculated.

61
 Figure 29: Piston velocity.

Once the displacement has been calculated, the velocity can be calculated when combined with
engine speed.

Figure 30: Piston acceleration.

The final stage in the series of differential equations is the piston acceleration which is an
important parameter as it allows the forces exerted on the piston, by the other engine dynamics,
to be calculated.

62
 Figure 31: Comparison of the primary and secondary forces.

The forces expected on the piston are made up of a primary force, from the reciprocating
movement of the piston, and a secondary force, from the rotating movement of the crank shaft.
Both these forces can be thought of as independent forces although they both act on the piston.

Figure 32: Combination of the primary and secondary forces due to the masses.

The combination of the primary and secondary forces give a non-uniform profile due to the
secondary force occurring at twice the frequency of the primary.

63
 Figure 33: Gas force applied onto the piston.

The gas force is a critical factor in the forces exerted onto the engine internals and is calculated
by the combustion model, although data can be acquired from the GT-Power model.

Figure 34: Analysis of the effects of force on same phase firing orders.

In order to analysis the total force acting on the internals, the gas force and the force to the
internal masses are combined. As the two pistons are out of line and can fire either in the same
phase or out of phase, it was deemed a worthwhile test to calculate the overall force that can
be exerted on the internals if the firing orders wherein phase and out of phase.

64
 Figure 35: Analysis of the effects of force on out of phase firing orders.

The analysis of the both the in phase and the out of phase firing orders showed that a firing
order 360 degrees out of phase reduced the overall force by a significant amount.

Figure 36: Analysis of the total force over the RPM range.

Once the forces and been calculated and the fire orders decided then the total forces could be
calculated over the RPM range. The 3 RPMs chosen to analyse the total forces were 4,000
RPM as this was ideal, 9,000 RPM as this was deemed optimal for peak torque, and 12,000
RPM as this was near the top of the RPM range.

65
 3.3. GT–Power model
The GT-Power model has been developed alongside a combustion model and analysis of the
engine dynamics, which has furthered the understanding of the engine. The model developed
using GT-Power was used to analyse the engine and continue its development.

The combustion object selected for the GT-Power model was the ‘EngCylSIWiebe’ model.
The object imposes a burn rate using the Wiebe function in order to approximate the shape, for
a typical combustion profile, of the SI burn rate. The model allows for implementing a
‘reasonable’ burn rate if in-cylinder pressure data is not available, which it was not for this
project, [12].

The parameters that were utilised in order to calculate the burn rate of the GT-Power Wiebe
model were, [12]; -

-Anchor Angle (AA), which is the anchoring angle of the Wiebe curve. The AA is the
number of crank angles between the TDC and 50% MFB.
-Duration (D), is the number of crank angles the combustion curve takes to get from 1
0% MFB to 90% MFB.
-Wiebe exponent (E), which is typically a default value of 2.
-Fraction of Fuel Burned (CE), which is the percentage of the total fuel burned in the
combustion event.
-Burned Fuel Percentage Anchor Angle (BM), which is 50% MFB and used as an
anchor point.
-Burned Fuel Percentage at Duration Start (BS), which is 10% MFB and is used to
define the start of combustion.
-Burned Fuel Percentage at Duration End (BE), which is 90% MFB and is used to
define the end of combustion.

The Wiebe model in the Combustion object calculates the constants; –

- BMC = - ln (1-BM), which is the burned midpoint constant. Equation 100

- BSC = - ln (1-BS), which is the burned start constant. Equation 971

- BEC = - ln (1-BE), which is the burned end constant. Equation 982

−(𝑬𝑬+𝟏𝟏)
𝑫𝑫
- 𝑾𝑾𝑾𝑾 = � 𝟏𝟏 𝟏𝟏 � , which is the Wiebe constant. Equation 993
𝑩𝑩𝑩𝑩𝑩𝑩(𝑬𝑬+𝟏𝟏) −𝑩𝑩𝑩𝑩𝑩𝑩(𝑬𝑬+𝟏𝟏)

𝟏𝟏
(𝑫𝑫)(𝑩𝑩𝑩𝑩𝑩𝑩)(𝑬𝑬+𝟏𝟏)
- 𝑺𝑺𝑺𝑺𝑺𝑺 = 𝑨𝑨𝑨𝑨 − 𝟏𝟏 𝟏𝟏 , which is the start of combustion Equation 1004
𝑩𝑩𝑩𝑩𝑩𝑩(𝑬𝑬+𝟏𝟏) −𝑩𝑩𝑩𝑩𝑩𝑩(𝑬𝑬+𝟏𝟏)

Once these constants have been calculated, then the cumulative burn rate is calculated
normalised to 1.0, with the combustion starting at 0.0 and is finished when the ‘fraction of fuel
burned attribute’ is calculated to be 1.0. The combustion objects calculates the Wiebe function
using the equation, [12] –
(𝑬𝑬+𝟏𝟏)
− 𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝑪(𝜽𝜽) = (𝑪𝑪𝑪𝑪) �𝟏𝟏 − 𝒆𝒆−(𝑾𝑾𝑾𝑾)(𝜽𝜽−𝑺𝑺𝑺𝑺𝑺𝑺) �

Equation 1015

66
With the architecture of the Wiebe function described in the combustion literature review
clearly visible, in the equation above.

Figure 37: The original engine model setup used in version GT-Power 6.1. The work was done carried out
at Oxford Brookes Universities by the students there.

This is a simplified model with a lot of assumptions that was re-created in GT-Power V7.1. A
lot of the features had changed but the parameters were given in spreadsheets and where input
into the model.

Figure 38: Original model of the engine used to analyse the changes proposed for the engine. In GT-
Power V 7.1. This was the original starting point of the project.

This was the first model constructed by the Formula Student V-Twin powertrain team in
version GT-Power V7.1.

67
 Figure 39: GT-Power model development flow chart

As seen above, the GT-Power model was started with the combustion model, shown by the
pistons. The crank case was then added with the engine dynamics. After the ports where added.
The ports had the valve train built into them. After the exhaust was added which discharges to
the atmosphere. Finally the inlet manifold was added, which has the injectors and as an inlet to
the open air. The final model can be seen below.

Figure 40: The final setup for the engine model in version GT-Power V2016.

68
The majority of parameters within a GT-Power model are set to a default setting ‘def’ in order
to give a generic engine model. If a more accurate model is wanted, or a better combustion
model is needed, then some of the parameters need changing. The more parameters changed
from ‘def’ to the optimal value, the more accurate the model will be for that parameter, and the
closer to the desired outcome. Most of the values acquired were a result a combination of the
work done on the combustion and engine dynamic models and work done on the background
of the engine.

The parameters of the GT-Power model can be found in GT-Power model development section
of the Appendix.

Figure 41: A representation the sweep function on GT-Power.

Figure 41: A representation the sweep function on GT-Power. Shows an example of the
optimisation of Camshaft Anchor angle on the GT-Power Optimiser. The optimisation process
used a sweep function that rans the model through different iterations of the same parameter
within a range of values that the user set. When analysing the valve timing,

The valve sweep of the valve train was done using a simplified model to make the process
quicker. The process was done with a sweep with steps across the full intake and exhaust
strokes. Using a simplified model does not give 100% accurate results but allows the full extent
of the area of max performance to be analysed more quickly, as the simplified model is far
quicker to simulate. A full model was used to identify the best possible valve timing setup for
the engine. Doing a full sweep of the strokes is also unrealistic as the valve would hit the piston
during parts of the stroke. A further analysis of the engine model will be carried out to verify
the crank angles at which either of the valve set will hit the pistons. Average torque analysis
was done between the results over the range of RPMs of the valve sweep. This was done to
show which had the smoothest torque curve.

69
 3.4. Valvetrain methodology

Figure 42: Comparison of Exhaust and Intake valve opening profiles Vs Displacement Volume, showing
original valve timings and areas of importance such as valve overlap.

The start of this process was to analyse the current valve train, by looking at the cam lobe shape
and design. The initial analysis of the cam lobe shape/cam profiles, shown in Figure 42:
Comparison of Exhaust and Intake valve opening profiles Vs Displacement Volume2 indicates
that it is set up for high engine revolutions. The valve overlap at 1mm of valve lift is over
50mm which is quite high. The EVO is almost at two thirds of the combustion stroke,
happening at 489 degrees Crank Angle (CA), showing that some of the combustion is being
compromised to get the maximum amount of exhaust gases out of the exhaust valve. The
majority of the usable combustion will have happened by the point of EVO. Advancing the
EVO has allowed for the IVO to be advanced, with the IVO at 1mm happening at 688 degrees
CA, and the crossover of the valve overlap happening over 10 degrees into the exhaust stroke.
This advancing of the IVO, and having the overlap occurring within the exhaust stroke, will
allow the maximum amount of exhaust gases to be forced out of the cylinder. The intake valve
is open for over 250 degrees CA, which is almost a third of the total CA degrees for the stroke.
The large opening CA range of the intake valve shows that the engine is set up for high RPMs,
as there is a compromise from air escaping back from either blow-back at the valve overlap or
air escaping back into the intake on the compression stroke, but both will help maximise the
amount of time available for air to get into the engine. Although there will be a slight loss in
performance, it will be compensated for by the number of power strokes happening within a
given time for the higher RPM of the engine.

70
 3.5. Manifold methodology

The CFD was done in SolidWorks FLowSim which is an integrated CFD package within
SolidWorks software. It uses the models constructed within the software in order to create the
flow body [1, 2, 7, 8].

3.5.1. Intake CFD

In order to calculate the mass flow rate for any CFD simulations, the maximum mass flow rate
into the system was deemed to be a critical factor. This is because the restrictor has an effect
on the amount of air the system can intake and causes a significant restriction on the system,
hence it is implemented by FASE via the rules to restrict engine power. The current restrictor
is an aluminium tube with the throttle inside the assembly. The current diameter is 19mm to
ensure that it is within the 20mm restrictor limit from the FASE formula student rules. The
1mm margin allows for tolerances in machining and in the official’s measuring equipment.

Figure 43: CAD of the AT-Power FSAE Throttle Part number from the FS servers FS08-02-520 (FS
CAD) [103]

To assess the restrictor, the mass flow rate of air through the system needs to be calculated.
This also helped in the construction of the combustion model in excel, and was done using
the equation –
𝛾𝛾+1
𝐴𝐴𝑡𝑡ℎ𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 × 𝑃𝑃 𝛾𝛾 𝛾𝛾 + 1 −2(𝛾𝛾−1)
𝑚𝑚̇𝑎𝑎𝑎𝑎𝑎𝑎 = ×� ×� �
√𝑇𝑇 𝑅𝑅 2
Equation 72

Which gave –

71
 1.4+1
4.57 × 10−3 × 101325 1.4 1.4 + 1 −2(1.4−1)
𝑚𝑚̇𝑎𝑎𝑎𝑎𝑎𝑎 = ×� ×� � = 2.0379 𝑘𝑘𝑘𝑘/𝑠𝑠
√300 287 2

Equation 72

For this calculation the air entering the restrictor was deemed to be at ambient conditions of
300K temperature and 1 bar pressure. The value of 2.0379 kg/s is the theoretical upper limit
and will be hard to achieve in reality due to the boundary layer between the restrictor surface
and the 19mm diameter throat [60].

Once the intake mass flow rate had been calculated then the intake runner lengths were
calculated in Excel and verified in GT-Power, the equation used was –

955 × 𝑐𝑐 2 𝐴𝐴
𝐿𝐿 = � � ×� �
𝐾𝐾 × 𝑁𝑁 𝑉𝑉𝑒𝑒𝑒𝑒𝑒𝑒

Equation 81

The results can be seen in the manifold results section.

72
 3.5.2. Exhaust CFD

The initial conditions for the exhaust CFD were estimated values that were taken from the
initial GT-Power model. The values used were 1 bar pressure and 1000K temperature, which
give a speed of sound of 633.88 m/s. The maximum theoretical flow through the port was also
calculated in order to give the CFD an initial mass flow rate into the system. The choked flow
limit of the exhaust port had to be assessed using the equation [68] –
𝛾𝛾+1
𝐴𝐴𝑡𝑡ℎ𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 × 𝑃𝑃 𝛾𝛾 𝛾𝛾 + 1 −2(𝛾𝛾−1)
𝑚𝑚̇𝑎𝑎𝑎𝑎𝑎𝑎 = ×� ×� �
√𝑇𝑇 𝑅𝑅 2
Equation 72

which gave –

1.4+1
3.19 × 10−3 × 101325 1.4 1.4 + 1 −2(1.4−1)
𝑚𝑚̇𝑎𝑎𝑎𝑎𝑎𝑎 = × � ×� � = 0.7792 𝑘𝑘𝑘𝑘/𝑠𝑠
√1000 287 2

Once the mass flow rate and initial gas conditions were input into the CAD and the simulations
run, it was an iterative process of comparing the results with the exhaust pressure traces from
GT-Power in order to calculate the optimal exhaust length. The initial exhaust and inlet runner
lengths were calculated with the equation –

955 × 𝑐𝑐 2 𝐴𝐴
𝐿𝐿 = � � ×� �
𝐾𝐾 × 𝑁𝑁 𝑉𝑉𝑒𝑒𝑒𝑒𝑒𝑒

Equation 81

The results can be seen in the manifold results section.

Figure 44: Pressure trace from the inlet runner simulated in GT-Power.

The GT-Power simulations and CFD were done alongside an Excel spreadsheet in order to
calculate the exhaust length that were then verified in the GT-Power model. Figure 44: Pressure
trace from the inlet runner simulated in GT-Power. shows the pressure trace from the inlet runner
simulated in GT-Power. The pressure trace shows the reciprocating pressure waves used in
intake design to maximise the RAM air effect that is used in manifold design.

73
 Chapter 4. Results and Discussion

74
 4. Results
4.1. Combustion model

Mass Fraction Burn (MFB) and Heat Release Rates (HRR) are two important parameters to
consider in engine development and in understanding engine behaviour. Both parameters relate
to how much of the charge is burned within the cylinder at any given time. One way of
measuring the HRR and the MFB is using the Wiebe function. This is the base model used
within the combustion model of GT-power. HRR data normally allows for a MFB profile to be
calculated for further analysis of the combustion event, [21].

Figure 45: Heat release per crank angle

The heat release rate per crank angle profile, seen in Figure 45: Heat release per crank angle5,
is similar to a MFB profile and the two are related as seen in Figure 46: Mass Fraction Burned
Vs Crank angle6.

75
 Figure 46: Mass Fraction Burned Vs Crank angle.

The start of the process was to find the volume of each cylinder using the column equation and
the graph from the Excel spreadsheet, seen in Figure 47: Displacement per cylinder of the V-
Twin including clearance volume.

Figure 47: Displacement per cylinder of the V-Twin including clearance volume

76
The equation employed in this calculation uses the individual gas constant R for air, which is
287J/Kg K. P0 and T0 are both reference parameters taken at ambient conditions so room
temperature and atmospheric pressure were used, and the volume V is Vc + Vd. Thus, the
equation is [1, 7, 41, 102]:

𝑝𝑝0 (𝑉𝑉𝑐𝑐 + 𝑉𝑉𝑑𝑑 ) 1 × 105 × ((26.2009 + 300) × 10−6 )

𝑚𝑚 = = = 0.3789 × 10−3 𝐾𝐾𝐾𝐾
𝑅𝑅𝑇𝑇0 287 × 300
Equation 72

Once the mass of air has been found, the mass of fuel can be worked out. For a gasoline engine
the stoichiometric air-fuel ratio is 14:7 so the mass of fuel can be worked out using the equation
[1, 7, 41, 102]:
𝛼𝛼 𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑎𝑎𝑎𝑎𝑎𝑎
𝜆𝜆 = 𝜙𝜙 −1 = ⟹ 𝛼𝛼 = = 14.7
𝛼𝛼𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠ℎ𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓
Equation 102

The equation can be rearranged for mass of fuel, which can be calculated thus [1, 7, 41, 102]:

𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑎𝑎𝑎𝑎𝑎𝑎 0.3789 × 10−3 𝐾𝐾𝐾𝐾

𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 = = = 0.0264 × 10−3 𝐾𝐾𝐾𝐾
14.7 14.7
Equation 8

Once the mass of fuel has been calculated then then the heat release can be calculated using
the equation [1, 7, 41, 102]:

𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑎𝑎𝑎𝑎𝑎𝑎 0.3789 × 10−3

𝑄𝑄𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶 = 𝑚𝑚𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 ℎ𝐿𝐿𝐿𝐿𝐿𝐿 = ℎ𝐿𝐿𝐿𝐿𝐿𝐿 = × 46 × 106 = 1185.6735𝐽𝐽
𝐴𝐴𝐴𝐴𝐴𝐴 14.7
Equation 103

where hLHV for gasoline equals 46 MJ/kg [1, 7, 41, 102]. The assumptions that were made in
order to calculate this number were that the air was at ambient conditions of 1 bar pressure and
300 degrees kelvin, the AFR was assumed to be Stoichiometric at 14.7:1, and the Lower
heating value was assumed to be 46 MJ/kg.

77
 Figure 48: Temperature comparison of the Wiebe and Otto cycles calculated using the Excel template.

The analysis of the combustion profiles of the two different methods clearly shows the
difference between them. Figure 48: Temperature comparison of the Wiebe and Otto cycles8
shows the heat increase of the Otto cycle is instantaneous and is over 1400 Kelvin greater than
on the Wiebe method. The Otto cycle shows a 30% increase in temperature over the Wiebe
method, which is unlikely. The Wiebe function shows a more realistic combustion profile and
demonstrates the problems with the assumptions made when calculating the Otto cycle.

78
 Figure 49: Pressure comparison of the Wiebe and Otto cycles, calculated using the Excel Template.

The Otto cycle in Figure 49: Pressure comparison of the Wiebe and Otto cycles9 shows a
pressure increase of 140 Bar higher than the Wiebe cycle, which again is deemed unrealistic.
The maximum pressure of the Wiebe function is 80 Bar, when combined with the 40% thermal
and 97% mechanical efficiencies normally seen in internal combustion engines a realistic
figure of 25-30 Bar is seen, which is to be expected from a high-performance race engine. The
overall energy efficiency of the system for the heat work can be found using the equation [1,
7, 41, 102] –

𝑊𝑊𝑛𝑛𝑛𝑛𝑛𝑛 1 1
𝜂𝜂 ≡ =1− = 1− = 0.6353
𝑄𝑄𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 𝑟𝑟 ϒ−1 12.451.4−1
Equation 12
This gives a total of 63.63% thermal efficiency which is an unrealistic value and far too high,
and is attributed to not taking into account other losses in the system, such as thermal loss to
the wall, unburnt fuel, pumping losses, loss of pressure round the piston ring and engine
friction. Most modern car engines have a thermal efficiency of around 35%, with F1 cars
having efficiencies of around 50% which are the highest seen in internal combustion engines
designed for cars. Such efficiencies are achieved by very careful control of the combustion
event, special lubricants and heat resistant coatings for the engine and using turbochargers and
an energy recovery system connected to the exhaust and gear box, which are well out of the
scope of this project in terms of time and budget. For this application of engine the overall
efficiency will be around 35% [1, 7, 41, 102].

79
 4.1.1. Conclusions
The analysis of the combustion model used in the GT-power model made assumptions
regarding the parameters a and m, as they needed to be consistent with the Wiebe function data
𝑑𝑑𝑑𝑑
represented in the dθ graph. This can only be done with experimentation data and not
theoretical data. Values of a=5 and m=3 were used in the first interpretations of the GT-power
model. Any future experimental data will allow the model to be made more accurate, as the a
and m values can be matched to the unique combustion profile of each individual engine [1, 7,
21, 41, 102]. The analysis of the final GT-Power model can be seen in section 4.4. GT-Power
model results.

80
 4.2. Dynamics model data and discussion

The first step of the analysis of the internal components is to check the current designs. One
parameter of the engine that needs to be assessed is piston speed as this has major implications
for the rest of the engine, the ideal speed being 30 m/s. An Excel spreadsheet was constructed
to run sweeps of the con-rod and crankshaft lengths and asses the effects they have on the
internals. An Excel spreadsheet was deemed the best method as the calculations have already
been assessed in the literature review. The analysis was done at 9,000rpm as this was where
peak torque would be ideal located and thus was used as an anchor.

Figure 50: Piston velocities by different con rod lengths.

The con-rod length makes a difference to piston speed. Step changes, during the sweep, of
around 25 mm, which gave differences of over 20 m/s.

81
 Figure 51: Piston velocities by different crank shaft offset lengths.

The crank offset had a limited impact upon piston speed. With sweeps of 5 mm the impact was
only a few m/s between the longest and the shortest crank offset.

Figure 52: Piston acceleration by different con rod lengths.

82
The con-rod has a big impact on the acceleration in terms of the forces applied to internals with
the same 20mm sweep. The noticeable difference with the sweep is the shape. A smoother
shape is preferable as it will aid in flow in and out of the cylinder.

Figure 53: Piston acceleration by different crank shaft offset lengths.

The crankshaft off-set makes a small difference in terms of acceleration over the same 10mm
sweep step, but has an impact on the shape of the acceleration profile. Again, the lowest peak
with the smoothest profile is the optimal shape.

Section Summary
Reducing the peak acceleration of the piston will result in more time for gas exchange. It will
also result in less force acting on the engine internals as F=ma, so reducing the acceleration
will reduce the force, given that the mass is staying the same. The maximum piston speed has
a lot of knock-on effects, and is ideally limited to around 30 m/s as this limits the forces exerted
on the internal components. The internal dynamics are directly related to the instantaneous
piston speed. When the lengths were analysed the original parameters of crank offset of
45.13mm and the con-rod length of 103mm were found to maintain a max piston speed of
30m/s, whist having the smoothest acceleration profile with the lowest peak.

83
Assessment of internal forces

The piston is an NME-Cosworth PA2622 with a bore of 92mm. The con-rod is an NME 103mm
Con-rod sourced from the Nicholson McLaren pattern for the Cosworth XB part, made by
Farndon Components. Both are made from steel. A table of the original engine specs can be
found in Table 1.

material property Con-rod Piston

Volume (m3) 1.2e-005 1.097e-004

Area (m2) 0.026 0.038

Mass (Kg) 0.418 0.297

Density (Kg/m3) 8410 2710

Table 3: Material Properties for the con-rod and piston

The starting point to analysing the engine internals starts with the piston characteristics

Piston displacement –
1
𝑠𝑠 𝑠𝑠 2 2
𝑋𝑋 = cos 𝜃𝜃 + 𝐿𝐿 �1 − � sin 𝜃𝜃� �
2 2𝐿𝐿
Equation 28
Piston velocity –

𝑠𝑠 𝜀𝜀
𝑋𝑋̇ = − 𝜔𝜔 �sin 𝜃𝜃 + sin 2𝜃𝜃�
2 2
Equation 29
Piston acceleration –
𝑠𝑠
𝑋𝑋̈ = − 𝜔𝜔2 (cos 𝜃𝜃 + 𝜀𝜀 cos 2𝜃𝜃)
2
Equation 35
Although to start the analysis of the engine internals, as per the literature review, the masses
have to be split up into the reciprocating and the revolving masses. This was done within
CATIA V5R20 as the material properties were added to the components, which showed where
the centre of mass was for the con-rod seen in Figure 54: Analysis of the con-rods centre of
mass using CATIA R5V24. and used the equation [1, 7, 38, 39, 42, 44, 45] –

𝐿𝐿𝑏𝑏 𝐿𝐿𝑎𝑎
𝑚𝑚𝐴𝐴 = 𝑚𝑚𝑐𝑐 𝐚𝐚𝐚𝐚𝐚𝐚 𝑚𝑚𝐵𝐵 = 𝑚𝑚𝑐𝑐 + 𝑚𝑚𝑝𝑝
𝐿𝐿𝑎𝑎 + 𝐿𝐿𝑏𝑏 𝐿𝐿𝑎𝑎 + 𝐿𝐿𝑏𝑏
Equation 35

84
 Figure 54: Analysis of the con-rods centre of mass using CATIA R5V24.

Once the Centre of Mass has been found the model of the con-rod can be split into two sections:
the reciprocating and rotating masses.

A) Showing the con-rod section added to the rotating mass.

B) Showing the con-rod section added to the reciprocating mass
Figure 55: The two sections of con-rod for each assembly of the reciprocating and the rotating masses.

The con-rod was split into the two sections to analyse each component of the mass, as part of
the mass affects the rotating and part effects the reciprocating mass. Figure 55: The two sections
of con-rod for each assembly of the reciprocating and the rotating masses.5 shows the split
con-rod masses and was done in order to add the optimal masses to the rotating show by A,
and reciprocating mass, shown by B, that are used to analyse the engine dynamics.

85
 Figure 56: CATIA Piston Assembly to analyse the mass.

The small end of the con-rod together with the piston assembly makes up the reciprocating
mass. The piston assembly is made up of the piston, small end of the con-rod, gudgeon pin,
gudgeon pin spacer and the various piston rings and circlips to hold the assembly together. The
reciprocating mass assembly has a weight of 0.567kg which, when factored into the analysis
of the piston characteristics, allow the forces on the engine to be calculated.

The big end of the con-rod is added to the rotating mass, along with the big end bearings, con-
rod bolts, and the crankshaft. The crankshaft was cut off at the axis to analyse the weight of all
the components above the axis, so the amount of counterweight could be ascertained. The
analysis found that the counter balance weight recommended by previous work effectively
balanced the rotating mass and will be discussed further on in the project.

86
 4.2.1. Forces within the engine

The first step in analysing the engine was to calculate the piston velocities. These were found
to reach peaks of just less than 30 m/s at 12k RPM, the engines redline, which is the highest
engine speed and therefore is the RPM where peak piston velocity should occur, seen in Figure
57: Piston velocity at 12.5k RPM acquired from Excel model.

Figure 57: Piston velocity at 12.5k RPM acquired from Excel model.

There are two pistons in the engine, each of which have forces associated with them, and these
forces are out of phase. The net effect of the combined forces has to be calculated with this in
mind. The pistons are 700 CA difference between the two cylinders and therefore the firing
orders are 700 CA out of phase.

87
 Figure 58: Analysis of both piston velocities.

The piston is subject to accelerations of up to 27,800 g in the positive direction and 43,400 g
in the negative direction, seen in Figure 59: Analysis of piston acceleration at 12k RPM..

Figure 59: Analysis of piston acceleration at 12k RPM.

88
The rotation of the crank shaft and the swinging of the con-rod gives the piston acceleration
profile seen in the piston acceleration graph and accounts for the sharp troughs. The primary
force associated with the reciprocating mass has peaks at around 20,000 Newtons, while the
secondary force peaks at around 4,500 Newtons. The primary forces are due to movement of
the piston mass, and the secondary force is a product of changing acceleration vectors through
the stroke. The primary and secondary forces will be felt as a vibrational force within the
engine. Vibrations reduce the smoothness of the engine as well as increasing fuel consumption
and reducing power due to the wasted energy. If the vibration are too excessive then the forces
will start to wear the internal components and can even cause catastrophic failures within the
engine.

Figure 60: Primary and Secondary reciprocating forces

The secondary force is roughly 22% of the total force and so cannot be ignored. The combined
force seen in Figure 60: Primary and Secondary reciprocating forces60, which peaks of 25,000
Newtons in the negative direction.

89
 Figure 61: Total reciprocating forces over a full 4 Stroke cycle.

The primary and the secondary forces due to the reciprocating masses have been summed. The
resultant force has reduced peaks in the positive direction of 1,600, but the peaks in the negative
direction are around 25,000, as seen in Figure 61: Total reciprocating forces over a full 4 Stroke
cycle.. This is due to the forces cancelling each other out in the positive direction whereas in
the negative direction the forces combine resulting in a larger peak.

4.2.2. Balancing rotating masses

Mass A is the mass rotating around the crankshaft, which creates a force that acts radially away
from the axis of rotation of the crankshaft. This force can cause wear and damage to the internal
components of the engine such as the bearing and crankshaft. In order to counter this force
another weight is positioned the opposite side of the rotational axis in order to counteract this
force as much as possible, preferably negating it altogether. Using equation 47, discussed in
the dynamics section of the literature review, to balance the rotating mass, it can be found the
force needed to be balanced at 12000 RPM is, [1, 2, 31, 33] –

𝐹𝐹𝐵𝐵 = 𝑚𝑚𝐴𝐴 (𝑟𝑟𝑟𝑟2 ) → (0.1224 × 1256.6372 ) × 1.48 = 286063 𝑁𝑁

Equation 467
This give a force to be balance of 286063 N or 286kN.

90
 Figure 62: Crank offset above the crank shaft axis.

The crankshaft offset was modelled in CATIA which gave an initial weight and offset for the
crankshaft seen in Figure 62: Crank offset above the crank shaft axis.. The next step was then
to add all the other components that make up the bottom end of the con-rod, including the
model of the cut off big end.

Figure 63: Rotating mass assembly above the axis of the crank shaft.

The model of the total mass that is connected to the crankshaft offset seen in Figure 63:
Rotating mass assembly above the axis of the crank shaft.. This is the mass that needs to be
balanced by a crankshaft counter-weight. The components that make up this assembly are the
big end bearings, the big end and the bolts. These are all taken into account in the mass analysis
with a total of 1.48 Kg.

91
 A) Showing the Properties of the unbalanced rotating mass

B) Showing the geometry and CoM of the unbalanced rotating mass

Figure 64: Total mass of rotating assembly with all the components.

CATIA can be used to assess the Centre of Mass in the model. Adding weight to the assembly
below the revolution axis will move the Centre of Mass towards the axis shown in Figure 64:
Total mass of rotating assembly with all the components.. Once the appropriate Gx, Gy or Gz

92
coordinate, depending on the model orientation, is roughly equal to 0 then the mass will be
balanced around the central rotating point, as the model axis is set on the rotational axis of the
crankshaft.

A) Counter weight for the rotating mass

B Size of counter weight for rotating mass

Figure 65: Analysis of counterweight in CATIA V5R20.

CATIA gives a centre of mass of the component when the material properties are input, which
allows the designs to be analysed quickly and accurately seen in Figure 65: Analysis of
counterweight in CATIA V5R205. To minimise the mass required for the counter-weight, it is
positioned away from the axis. This is because, for a given weight, the further from axis a
given weight is, the larger the effect/counter force is on the other components. The counter-
weight is positioned as far from the axis as possible, whilst retaining structural integrity of the
whole assembly, to allow for the minimum counter-weight to be used, therefore minimising
the weight of the whole assembly. This makes the engine lighter and minimises the forces
acting on different components, which reduces fatigue and wear within the engine.

93
 A) Rotating mass assembly with material properties.

B) Isometric view of rotating mass assembly with the material properties.

Figure 66: Finished rotating assembly with counter weight added in.

It is common practice in the automotive industry to over-balance the crankshaft to compensate

for the gas loads. However, the gas loads are not even, and the over-balancing will cause
excessive force on the engine components for more than the short period of time the gas force
is acting on the piston which is addressed in section later section.

The total rotation inertia for the rotational mass and the counter balance was shown to be in
balance as the rotational mass had a centre of mass 0.1224 meters for the rotational axis while
the counter balance mass had its centre of mass 0.1725 meters for the rotational mass and they
had a total mass of 1.48 Kg and 1.05 Kg respectively as shown below –

94
For the rotating mass –

𝐹𝐹𝐵𝐵 = 𝑚𝑚𝐴𝐴 (𝑟𝑟𝑟𝑟2 ) → (0.1224 × 1256.6372 ) × 1.48 = 286063 𝑁𝑁

Equation 47

For the counter balance mass –

𝐹𝐹𝐵𝐵𝐵𝐵𝐵𝐵 = 𝑚𝑚𝑐𝑐𝑐𝑐−𝐵𝐵 (𝑅𝑅𝑅𝑅2 ) → (0.1725 × 1256.6372 ) × 1.05 = 286063 𝑁𝑁

Equation 48

One argument for over-balancing the crank to account for the gas force is to reduce the pressure
on the top half of the bearing, but a well-designed oil system will create an oil pocket between
the bearing and crankshaft which will dissipate this force. The gas loads are well in excess of
the compensating force delivered by an excessive counter balance, which is therefore not
desirable due to the extended strain exerted on the crankshaft from overbalancing.

Figure 67: Torque inertia of the combined reciprocating masses at 12k RPM.

Using the same Excel template, the torque inertia was assessed and the results implemented
into the GT-power model to improve the accuracy of the simulations.

95
Assessment of the combined forces
The Pressure trace was extracted from GT-Post and was input into the dynamics model, in
Excel, for post simulation analysis. The results, seen in Figure 68: Pressure plot from GT-Post
and plot of Gas force from Excel dynamics model, show the in-cylinder peak pressure rises to
over 80 bar, which equates to around 55,000 Newtons of force exerted on the piston at 9,000
RPM.

A) Pressure plot from GT-Post at 9000 RPM.

B) Excel dynamics model plot of the gas force at 9,000 RPM.

Figure 68: Pressure plot from GT-Post and plot of Gas force from Excel dynamics model.

96
Once the gas forces and inertia of the reciprocating masses are combined it is possible to
analyse the overall force acting on the crankshaft, as shown in Figure 69: Combined Gas forces
and inertia of reciprocating masses from the Dynamics model..

Figure 69: Combined Gas forces and inertia of reciprocating masses from the Dynamics model.

Figure 69: Combined Gas forces and inertia of reciprocating masses from the Dynamics
model.9 shows the forces from the reciprocating mass increases with engine speed. However,
when combined with the gas forces, the 9,000 RPM gas force has a proportionately larger
impact on the overall force as it makes up a much larger load than the equivalent at 12,000
RPM.

97
 A) Out of phase firing order loads.

B) Same phase firing order loads

Figure 70: A shows out of phase firing order loads. B shows same phase firing loads. Both at 12k RPM.

The combustion phasing needs to be analysed as this will dictate the gas loads applied to the
internal components at different crank angles. The firing orders are always 75 crank-angle
degrees out of phase due to the 75 Degree V on the engine, but the phase difference is a factor
of 180 degrees out, as that is the amount of rotation one of the 4 combustion phases requires.
Figure 70: shows the same-phase firing order exhibits a maximum load of 37,000 Newtons
and the out-of-phase has a maximum load of roughly 33,000 Newtons. Therefore to reduce
the loads on the internal components the 360 degrees out-of-phase firing order will be used.

98
4.2.3. Section Summary
The work carried out in developing the dynamics model for the V-twin engine has not only
allowed the current dynamics to be explored, and further understanding to be achieved. The
work carried out in the dynamics section helped develop the GT-Power model by aiding
detailed analysis of the components and their masses, and the data was input into the GT-Power
model where relevant.

4.3. Manifold design and analysis

The inlet manifold was analysed within the GT-Power model to assess the current designs. Due
to the iterative process of engine development, the GT-Power model was used to analyse the
initial calculations and allowed further analysis of any potential improvements that can be
demonstrated by the simulations.

4.3.1. Intake plenum size

Figure 71: CAD model of the original intake manifold

One restriction found with the original design was the restrictor was too small, the original
2.67 litre plenum was found to need an increase in volume to 4.12 litres. Combined with the
other modifications proposed, such as different runner lengths and exhaust length, this gave a
10BHP increase over the original model.

99
 4.3.2. Intake manifold

There were some other areas of the intake that could potentially be improved such as the shape.
The original had a flat bottom where as a rounded shape could give flow benefits as there are
less sharp corners within the system.

Figure 72: CAD representation of what a potential intake manifold could look like with the proposed
modifications.

100
4.3.3. Runner analysis

The equations for the intake were calculated at ambient conditions of 1 bar atmospheric
pressure and 300K room temperature, which gave a speed of sound of 344 m/s. The calculated
lengths of the inlet runner were analysed in GT-Power, with their harmonic orders as well, to
ascertain the optimal length for the runner. The results can be seen in Table 4.

Intake runner length in mm

Harmonic Order
Intake
1 1.5 2 2.5 3 3.5 4 4.5
4000 3572.74820 2381.83214 1786.37410 1429.09928 1190.91607 1020.78520 893.18705 793.94405
5000 2286.55885 1524.37257 1143.27942 914.62354 762.18628 653.30253 571.63971 508.12419
6000 1587.88809 1058.59206 793.94405 635.15524 529.29603 453.68231 396.97202 352.86402
7000 1166.61166 777.74111 583.30583 466.64466 388.87055 333.31762 291.65291 259.24704
8000 893.18705 595.45803 446.59353 357.27482 297.72902 255.19630 223.29676 198.48601
RPM
9000 705.72804 470.48536 352.86402 282.29122 235.24268 201.63658 176.43201 156.82845
10000 571.63971 381.09314 285.81986 228.65588 190.54657 163.32563 142.90993 127.03105
11000 472.42951 314.95301 236.21476 188.97181 157.47650 134.97986 118.10738 104.98434
12000 396.97202 264.64802 198.48601 158.78881 132.32401 113.42058 99.24301 88.21601
13000 338.24835 225.49890 169.12418 135.29934 112.74945 96.64239 84.56209 75.16630
Table 4: Intake runner length and harmonic order comparison.

Figure 73: Pressure trace from the inlet runner simulated in GT-Power. shows the pressure
trace from the inlet runner simulated in GT-Power, showing the reciprocating pressure waves
used in intake design to maximise the RAM air effect that is commonly used in manifold
design.

The method of calculating the runners can be seen in the manifold methodology.

Figure 73: Pressure trace from the inlet runner simulated in GT-Power.

GT-power optimiser was used to find the ideal runner length, as well as using an iterative
process of testing which looked at the pressure waves that were created for different runner
lengths. The ideal length was found to be 380 mm, but this is far too long because the runners
would have interfered with each other, as seen in Figure 74: Runner angle analysis and
curvature. A way of shortening the runner length was to taper the runner, as the tapering effect
gives the acoustic properties of a longer runner of constant diameter. This is because the
increase in diameter slows down the speed of sound within the runner and therefore slows down
the pressure waves. Thus, a shorter, tapered pipe will give a similar time for a pressure wave
to travel down it, as would a longer straight pipe.

101
The mass flow rate of 0.3109 kg/s was used as the outlet mass flow rate of the port.

Figure 74: Runner angle analysis and curvature.

Figure 74: Runner angle analysis and curvature shows that if the runners were to follow a
straight path from the inlet port they would intersect, which means they need to be curved to
avoid flow intersection and allow them to fit. Another consideration for curving them was in
order to fit the fuel injectors.

Figure 75: Length reducing expansion representation of the runner.

Figure 75: Length reducing expansion representation of the runner shows runners that have the
same acoustic properties, but the runners that are flared have a significantly shorter length.
The GT-Power model indicated that the flared runner length should be 333mm, compared a
straight runner length of 380mm.

102
4.3.6. Exhaust design
The equations for the exhaust were calculated using estimated values that were taken from the
initial GT-Power model. The values used were 1 bar pressure and 1000K temperature, which
give a speed of sound of 633.88 m/s. In table 5 the potential exhaust lengths depend on their
harmonic orders.

Exhaust length in mm
Harmonic Order
Exhaust
1 1.5 2 2.5 3 3.5 4 4.5
4000 4353.20326 2902.13550 2176.60163 1741.28130 1451.06775 1243.77236 1088.30081 967.37850
5000 2786.05008 1857.36672 1393.02504 1114.42003 928.68336 796.01431 696.51252 619.12224
6000 1934.75700 1289.83800 967.37850 773.90280 644.91900 552.78772 483.68925 429.94600
7000 1421.45412 947.63608 710.72706 568.58165 473.81804 406.12975 355.36353 315.87869
8000 1088.30081 725.53388 544.15041 435.32033 362.76694 310.94309 272.07520 241.84463
RPM
9000 859.89200 573.26133 429.94600 343.95680 286.63067 245.68343 214.97300 191.08711
10000 696.51252 464.34168 348.25626 278.60501 232.17084 199.00358 174.12813 154.78056
11000 575.63018 383.75346 287.81509 230.25207 191.87673 164.46577 143.90755 127.91782
12000 483.68925 322.45950 241.84463 193.47570 161.22975 138.19693 120.92231 107.48650
13000 412.13759 274.75839 206.06879 164.85503 137.37920 117.75360 103.03440 91.58613

Table 5: Table of engine speed Vs harmonic order to give exhaust lengths in mm

The method for calculating the exhaust length can be seen in the manifold methodology section,
previously. When comparing the results of the harmonic analysis of the exhaust length it is
necessary to ensure that the overall exhaust is sufficient to allow clearance from the chassis
and any critical components. This, however, this will affect the aerodynamics of the car and
thus have to be examined further but the aerodynamics and chassis developers. From the point
of view of this project, for engine performance, peak torque is delivered at 9000 RPM.
Therefore, the exhaust length should be 859.89mm which matches the 1st harmonic order. This
is long enough to exit the chassis and is a recommendation from this project that should be
further examined by the aerodynamics and chassis developers.

Furthermore, analysis of the flow through a mock-up of the exhaust system shows that there
could be some improvement to the original intended design of exhaust. The original exhaust
system design shows flow problems, highlighted by the flow separation from the surface,
shown in Figure 75.

Figure 76: Flow analysis of exhaust pressure distribution using CFD.

103
Figure 77: analysis of flow within the current and proposed exhaust system. shows the flow in
the original system has a lot of flow separation from the system wall, which causes pressure
build ups.

A) Current design pressure distribution. B) Proposed design pressure distribution.

Figure 77: analysis of flow within the current and proposed exhaust system.

A) Current design pressure distribution. B) Proposed design pressure distribution.

Figure 78: Side by side comparison of the pressure within the current exhaust and the original.

Figure 78: Side by side comparison of the pressure within the current exhaust and the original.
shows the smoother bends in the proposed design change allow a smoother run for the exhaust
gas to flow through the system and reduces the flow separation, resulting in less turbulent flow.

104
A) Current design velocity distribution. B) Proposed design velocity distribution.

Figure 79: Side by side comparison of the velocity within the current exhaust and the original.

Figure 79: Side by side comparison of the velocity within the current exhaust and the original.
shows the smoother flow through the proposed system, meaning reduced flow resistance,
resulting in higher velocities than in the original design. Higher flow velocity in the exhaust is
important because it reduces the pressure. Reducing the pressure creates a higher pressure
differential between the exhaust and the cylinder which improves the extraction of exhaust gas
within the cylinder, which can improve the combustion due to reduced exhaust gas left over in
the cylinder during the intake stroke.

Section Summary

The exhaust system needs to be of a length that would protrude into the air flow around the car
and therefore would interfere with the aerodynamics. Therefore, the chassis designers and
aerodynamicists will be given the optimal length but the final placement of the exhaust should
be left to the aero department. The exhaust length indicated by the GT-Power model is 850mm
long between flow splits, which is an increase from the 668mm in the original design.

Section summary

This project found that the inlet manifold design can be improved in a number of ways. An
increase in size of the plenum from 2.67 to 4.12 litres would give a 10BHP increase in power
performance over the original design. Giving the plenum a rounder shape could further increase
performance. Analysis of the runner design shows that the use of curved, flared runners also
improves performance by maximising the RAM air effect within the required dimensions. The
project also found that the exhaust design can also be improved. First, having a smoother curve
in the exhaust manifold reduces pressure in the system and improves gas flow. Secondly, by
matching the exhaust length to the appropriate harmonic order for 9000 RPM, which is the
RPM chosen for peak torque, performance is also improved. Taken together, these design

105
improvements could result in a giving a maximum of 6% increase in peak torque at 9000 rpm
and a 19% increase in peak torque at 12000 rpm.

4.3. V-twin engine performance and calibration

Valve train and control strategies

The sweeps for the valves were analysed via GT-power simulation and the results were input
into Excel spreadsheets.

Torque Intake Valve Anchor Crank Angle

Exhaust 360 405 450 495 540
Valve 180 -3.69 -3.59 -3.61 -3.94 -3.77
Anchor 225 8.09 12.73 13.08 8.69 2.17
Crank 270 13.29 16.51 15.53 9.66 -2.32
Angle 315 12.54 16.68 15.12 10.09 -2.24
360 0.97 -1.45 14.04 8.65 -1.45
Table 6: Max Torque Valve Anchor Sweep Table of Results.

Power Intake Valve Anchor Crank Angle

(HP)
Exhaust 360 405 450 495 540
Valve 180 -2.26 -2.36 -2.20 -2.21 -2.12
Anchor 225 11.36 18.41 22.21 14.64 -1.37
Crank 270 12.13 23.17 24.81 16.28 -1.92
Angle 315 7.30 19.26 22.25 13.87 -1.57
360 0.95 -1.22 15.48 8.47 -1.22
Table 7: Max Horse Power Valve Anchor Sweep Table of Results.

Figure 80: Max Power Valve sweep.

106
Figure 80: Max Power Valve sweep indicates the intake and exhaust valve timing that will give
max Horse Power. This area is 450-495 CA for the intake and 180-225 CA for the exhaust.

Figure 81: Max Torque Valve sweep.

The area of interest for the torque is 405-450 CA for the intake and 225-270 CA for the exhaust
as seen in Figure 81.

Figure 82: Average Horse Power Valve sweep

107
The average horse power valve sweep highlighted a specific area compared to the other sweeps.
The area was around the 450 CA intake and the 270 CA exhaust, as seen in figure 82.

Figure 83: Average Torque Valve Sweep

The average torque sweeps showed an area that was 450-495 CA of the intake and 225-270
CA of the exhaust, as seen in Figure 83.

Once the broad sweeps had been analysed then a more focused valve sweep was done to narrow
the search down to a specific crank angle. The final sweep was done over 405-495 CA intake
and 225-315 CA exhaust. The analysis was done at 9,000 RPM, as the chosen anchor point for
peak torque.

108
 Figure 84: Detailed intake valve Sweep

The detailed intake sweep shows the peak HP and Torque was achieved at 445 CAo.

Figure 85: Detailed exhaust valve Sweep

The detailed exhaust sweep shows that the peak power and Torque was achieved at 250 CAo.

109
 4.3.3. Valve train

The different sweeps brought up different areas of interest, depending on whether average or
max torque or horse power was looked at. The valve timings of 250 CA and 445 CA were
chosen as the optimal settings. The movement of the valve train anchor angle gave an increase
of 2% in peak torque across the midrange, with less than 1% decrease in performance above
11000 RPM and below 5000 RPM. Combined with the improvements proposed for the intake
and exhaust, this change in valve timings will further increase performance.

4.4. GT-Power model results.

This section discusses the engine performance and combustion characteristics expected
following implementation of the new manifold design and geometry. The GT-Power model
results show that the new manifold design significantly affects the engine performance and
combustion. Particularly, the volumetric efficiency and combustion duration are significantly
improved, which leads to the higher effective work output, indicate by IMEP.

Parameter Unite Original Model Final Model

Torque Nm 62.92 65.60
Engine Power Kw 59.31 61.83
IMEP Bar 14.79 15.35
Voleff % 85.05 88.96
Air Flow Rate Kg/hr 159.67 167.01
Peak Pressure Bar 78.37 78.12
0
MFB10 CA -2 -2.1
0
MFB50 CA 11.20 10.51
0
MFB90 CA 35.3 29.30
0
MFB10-90 CA 33.34 27.21

Table 8: Highlighted engine performance parameters

The figures in
Table 8: Highlighted engine performance parameters show the key engine performance
parameters that were used to compare the performance of the GT-Power models. The GT-
Power model development can be seen in the GT-Power section of the Appendix.

110
 4.4.1. In-cylinder Pressure

The original and the final models were compared at 9000 RPM, this being the RPM chosen for
peak torque, and allowing the two models to be easily compared and contrasted.

Figure 86: Original vs final GT-Power model in-cylinder pressure data.

The peak in-cylinder pressure of the final model is 78 Bar, which is an increase of 5 Bar over
the peak of 73 Bar in the original model. The IMEP of the final model is 15.34 Bar which is
higher than the original model’s IMEP of 14.79 Bar, showing the average in-cylinder pressure
in the final model is higher than that of the original. A reduced peak pressure should also
marginally reduce the wear on the internal components. The peak pressure difference between
the two was 7 Bar, this difference being in line with the power difference between the two
models.

111
 4.4.2. Pressure vs Volume diagram

Figure 87: Original and Final GT-Power model Pressure vs Volume.

The pressure-volume diagrams show very similar traces. However, when analysed the final
model shows a larger increase in pressure per change in volume when the piston is near TDC.
The trace shows that there is a larger increase in pressure towards the peak before the volume
starts to increase. After the peak pressure, the final model maintains its pressure per unit volume
better than the original. The maintenance of pressure would explain why the IMEP is clearly
higher in the final model, while having a smaller peak. The original model has an IMEP of
14.8 bar while the final model has an IMEP of 15.3 Bar.

112
 4.4.3. Mass Fraction Burned

Figure 88: Original GT-Power Mass Fraction Burned trace.

Figure 89: Final GT-Power model Mass Fraction Burned.

The final model’s MFB10-90, which represents combustion duration, was over 5 CA0 less than
the original’s 33.3 CAo. The shorter combustion duration of the final model would result in the
higher average in-cylinder pressures. Though the MFB10 and MFB50 were at similar CA in
both the original and final models, the final model’s MFB90 was over 60 sooner within the
cycle, showing that the second phase of combustion happened in a shorter time period in the
final model.

113
 4.4.4. Power performance Curve

Figure 90: Original GT-Power models power performance curve.

Figure 91: Final GT-Power models Power Performance curve.

The final model has a torque curve that shows a clear improvement at between 9,000 RPM to
10,000 RPM, although there was still a depression around 5,000 RPM. The idle of the engine
is 4,000 RPM so the depression at 5,000 RPM was not considered too detrimental to the
engine’s performance. However, the increase in torque after 9,000 RPM results in a more stable
and consistent power curve, as seen in Figure 91: Final GT-Power models Power Performance
curve.1. The power curves show that the final model shows significant improvement in engine
power performance. The drop off in power in the original model indicates a restriction caused
by the plenum being too small.

114
 Figure 92: Power Curve Comparison

The proposed improvements reduced performance by about 5% up to 7000 RPM, but thereafter
increased power performance, from a 6% increase in power performance up to a 19% increase
in performance at the top-end of the rev range. The loss of low-end power is worth the
significant improvement in the top-end power, as this engine is designed to be used on a
racetrack where low-end performance is not important, but top-end power is vital. The new
design shows particular improvement compared to the drop off in top-end power of the original
model. As seen in Figure 92: Power Curve Comparison2, the final design shows a smoother
power curve that doesn’t drop away as the original model’s power curve does. The smoother
power curve is due to combined effects of the changes in the manifolds and valve timings with
a more top-end of the engine’s RPM range focused optimisation. Figure 92: Power Curve
Comparison2 also shows the clear difference in the power curve from the simulations done in
the previous version of GT-Power, version V6.1.

The difference between the two power curves is in part due to a combination of a change in
plenum design, a change in intake design, improved exhaust manifold design and exhaust
length. It is also due to the change in valve timing anchor, as the valve train overlap is 2 degrees
more. The change in valve timing will allow more time for the scavenging effect and therefore
allow more clean air into the piston. This benefit is seen at higher RPM. The scavenging effect
would explain the increase in power performance at the top-end of the power curve when
approaching the 12,500 RPM redline.

4.4.5. Section Summary

The performance increase from the improvements to the engine proposed in this project were
clear. The area that showed a large improvement over the original model was the volumetric
efficiency, which is due to improved combustion, demonstrated by the MFB figures. The
overall power of the final model was higher across the rev range, apart from at 5,000 RPM,
with a peak 90 BHP in the final model, compared with 70 BHP in the original. An interesting
comparison between the work carried out in this project and the previous work on V6.1 is the
drop in the power curve from 7,000 RPM for all three models and again a slight drop at 10,000
RPM. This drop at 7,000 RPM and again at 10,000 RPM is likely to be caused by the restrictor
or the ports, as these were the only things which remain the same in each model.

115
 Chapter 5. Conclusions

116
5. Conclusions
This project looked into the development and optimisation of a bespoke Formula Student V-
twin engine based on 2 cylinders of a Cosworth DFV V8, to the point where it could be can be
approved for testing on a Dynamometer. As the engine has not yet been prepared for dyno
testing, there is no real time data available in order to validate the combustion and GT-Power
models, therefore the work done in this project is purely theoretical and must be validated when
the engine has been run and the results examined. However, the project has highlighted several
areas of design improvement which should significantly improve performance.

This project has shown great potential to improve the V-Twin’s performance for Formula
Student (FS) competitions in the future. The original target figure set by the engine design team
at the start of the project was a power output of 100 BHP. Upon analysing the basics of the
engine, and taking into account the rules set out by the FSAE FS event rules, it was concluded
that although 100 BHP could be achieved using turbochargers, in Naturally Aspirated form the
engine will be limited to about 96-97 BHP by the air restrictor required by the regulations.
This theoretical figure of 96-97 BHP is the upper limit and will be hard to achieve when the
car is on the track. The difference between the original model and the model incorporation the
improvements recommended in this project, in terms of peak pressure at 9000 RPM, was 4.6
Bar, which is expected from the power performance increase shown by the power curves.

In order to improve engine performance, this project investigated the deigns of both intake and
exhaust systems. Individual sections were looked at, including the restrictor, the intake plenum,
the runners and exhaust manifold. The project found that significant improvements in power
output where achievable through changes in the design of the inlet manifold, including plenum
size and shape, and length and shape of the runners, and the shape of the exhaust manifold and
pipe lengths. The later will require further consideration by aerodynamicists and chassis
designers as the final cars design is refined. Increasing the plenum size from 2.67 to 4.12 litres
gave a maximum of 6% increase at 9000 rpm and a 19% increase in performance at 12000 rpm.
The other changes indicated by the project are an increase in runner length from 333mm to
380mm, and an exhaust length increase from 668mm to 850mm.

The valve train was analysed using the GT-Power model that was developed alongside the
combustion and dynamics models. The results for the valve train show that the design is near
optimal. The cam anchor angle was the valve train parameter that was found to have the biggest
impact on the engine performance. This would, however, also depend on the ambient
conditions of the track and the engine map/spark timing chosen. The valve timing
recommended is exhaust CA of 250 and intake CA of 455, which is only a minor change from
the original exhaust CA of 255 and intake CA of 462. The movement of the valve train anchor
angle gave an improvement of 2% increase across the midrange with less than 1% decrease in
performance below 5000 RPM and above 11000 RPM. This is due to a combination of factors.
The increase in performance in the midrange is due to the increase in valve overlap of 2 degrees
and the resulting pressure differential. The decrease in performance is due to the reduced time
for the combustion event at low and very high RPMs resulting from the increased valve
overlap. The improvement delivered in the mid RPM range combined with the improvements
to the inlet and exhaust systems will deliver a more significant improvement in power
performance overall.

Incorporating all the design changes identified in this project delivers a theoretical peak power
performance of the engine of 90 BHP, this is a significant improvement on the original model’s
of 79 BHP. Similarly, the average power performance rose from 62 to 66 bhp. Furthermore,

117
the final model’s torque curve was smoother than the original model. The proposed
improvements to the engine showed a loss in performance of about 5%, between 4000 RPM
and 7000 RPM, which is a less important RPM range for a race car, but in the significant RPM
range for racing power performance was improved from 6% up to 19% towards the top-end of
the RPM range. This is significant improvement in engine design.

The original restrictor design is deemed satisfactory, as changes would result in only minor
improvements in performance while the current design ensures compliance with the rules, and
the current valve train and internal components were found to be satisfactory. However, new
hardware is recommended as a result of this project including a larger and more rounded
plenum, flared, curved runners and a more rounded exhaust manifold. Runner and exhaust pipe
lengths are also modified. The hardware changes recommended will deliver significant
improvements to power performance.

Overall the changes to hardware design and timings identified in this project should deliver
significant power performance improvements to the engine that was the subject of this study.
It is therefor recommended that they be incorporated into the design of the engine to be taken
forward to the next stage of development.

Further work
The next stage for development of this engine is to get it running and to validate the simulation
results from this project. This will allow final design improvements for the engine to be
optimised, within the rules, and to the drivers’ liking.

The exhaust must be designed in parallel with the chassis designers and aerodynamicists, and
cannot be completed until the designs can be validated with a running engine.

This project successfully identified scope for optimisation of a Naturally Aspirated (NA)
engine. However, a previous project found that Turbocharging the engine would deliver
significant gains to engine performance. Turbocharging could therefore be considered as part
of the future work for the engine.

118
6. References
[1] Heywood, J. B. Internal Combustion Engine Fundamentals. McGraw Hill, Automotive
Technology Series, 1988.

[2] Nigus, H. Kinematics and Load Formulation of Engine Crank Mechanism. Federal TVET
Institute, School of Mechanical Technology, Automotive Technology Department, Addis
Ababa, Ethiopia. Mechanics, Materials Science & Engineering, ISSN 2412-5954. 2015.

[3] Dr Timmins, S. Power Train Theory. Department of Mechanical Engineering. University

of Delaware. Available at http://www.me.udel.edu/meeg425/. 2013.

[4] Ranjbarkohan, M. Asadi, M, R. Rasekh, M. Hoseini, A, H. Kheiralipour, K. Kinematics

and kinetic analysis of the slider-crank mechanism in otto linear four cylinder Z24 engine.
Department of Mechanical Engineering, Islamic Azad University, Buinzahra branch, Qazwin,
Iran. Department of Agricultural Machinery, University of Mohaghegh Ardabili, Ardabil, Iran.
Mechanic Engineering in Megamotor Company, Tehran, Iran. Department of Mechanical
Engineering of Agricultural Machinery, University of Tehran, Karaj, Iran. Journal of
Mechanical Engineering Research Vol. 3(3), pp. 85-95. ISSN 2141 – 2383. 2011.

[5] Chatlatanagulchai, W. Rhienprayoon, S. Yaovaja, K. Wannatong, K. Air/Fuel Ratio

Control in Diesel-Dual-Fuel Engine by Varying Throttle, EGR Valve, and Total Fuel. Kasetsart
University, PTT, Thai-German Petroleum & Chemical, PTT. SAE Technical Paper 2010-01-
2200. 2010.

[6] Turner, J, W, T. Pearson, R, J. Bell, A. de Goede, S. Woolard, C. Iso-Stoichiometric Ternary

Blends of Gasoline, Ethanol and Methanol: Investigations into Exhaust Emissions, Blend
Properties and Octane Numbers. Lotus Engineering, Sasol Technology (Pty) Ltd, University
of Cape Town. SAE Technical Paper 2012-01-1586. 2012.

[7] Stone, R. Introduction to Internal Combustion Engines. Basingstoke: Macmillan, 1985.

[8] Wagner Roberto da Silva Trindade, Rogério Gonçalves dos, Combustion modelling applied
to the engines using 1D simulation code, 25th SAE Brasil International, Congress and Display,
Sao Paulo, Brasil, 2016.

[9] Liang, Y. Pope, S B. Pepiot, P. An adaptive methodology for the efficient implementation
of detailed chemistry in simulations of turbulent non-premixed Combustion. Sibley School of
Mechanical and Aerospace Engineering, Cornell University. Paper # 114TF-0378. 2015.

[10] Grill, M. Bargende, M. Rether, D. Schmid, A. Quasi-dimensional and Empirical Modeling

of Compression-Ignition Engine Combustion and Emissions. FKFS, IVK Univ. of Stuttgar.
SAE paper 2010-01-0151. 2010.

[11] Harish Kumar, R. Antony, A, J. Progressive Combustion in SI-Engines—Improved

Empirical Models for Simulating and Optimizing Engine Performance. Sri Siddhartha Institute
of Technology, Sahyadri Institute of Technology. SAE paper 2008-01-1630. 2008.

[12] Engine Performance Application Manual. GT-Suite. Version 7.5. Gamma Technologies.
PDF. www.gtisoft.com. 2014.

119
[13] Dr Hasse, C, F. Dr Wensing, M, E. Combustion modeling for virtual SI engine calibration
with the help of 0D/3D methods. Faculty of Mechanical Engineering, Process Engineering and
Energy Engineering. Technical University Bergakademie Freiberg. Von der Fakultät für
Maschinenbau,Verfahrens- und Energietechnik der Technischen Universität Bergakademie
Freiberg genehmigte. zur Erlangung des akademischen Grades Doktor-Ingenieur (Dr.-Ing.)
vorgelegt. Von Diplom-Ingenieur Sebastian Grasreiner aus Erfurt. 2012.

[14] Wang, S. Prucka, R. Zhu, Q. Clemson University. Prucka, M. Dourra, H. FCA US LLCA.
Real-Time Model for Spark Ignition Engine Combustion Phasing Prediction. SAE Technical
Paper 2016-01-0819. 2016.

[15] Siokos, K. He, Z. Prucka, R. Clemson University. Assessment of Model-Based Knock

Prediction Methods for Spark-Ignition Engines. SAE Technical Paper. 2017-01-0791. 2017.

[16] Klimstra, J. The Optimum Combustion Phasing Angle - A Convenient Engine Tuning
Criterion. Research Department N. V. Nederlandse Gasunie Groningen, The Netherlands. SAE
Technical Paper 852090. 1985.
[17] Salazar, F. internal combustion engine. Department of aerospace and mechanical
engineering. University of Notre Dame. 1998.

[18] The Calculation of Mass Fraction Burn of Ethanol-Gasoline Blended Fuels Using Single
and Two-Zone Models, Michigan Technological University, Yeliana. C, Cooney. J, Worm. J,
Naber. SAE 2008-01-0320, 2008

[19] Brown, B, R. Combustion Data Acquisition and Analysis. Final Year Project. Available
at https://www.catool.org/files/FinalProjectReport.pdf. Department of Aeronautical and
Automotive Engineering, Loughborough University. 2014.

[20] Yeliana, C. Cooney, J. Worm, D. Michalek, J. Naber. Wiebe Function Parameter

Determination For Mass Fraction Burn Calculation in a Ethanol-Gasoline Fuelled Engine,
Michigan Technological University, Journal of KONES Powertrain and Transport. 2008.

[21] Yeliana, C. Cooney, J. Worm, D. Michalek, J. Naber. Wiebe Function Parameter

Determination for Mass Fraction Burn Calculation in an Ethanol-Gasoline fuelled SI Engine.
Michigan Technological University Journal of KONES Powertrain and Transport, Vol. 15, No.
3. 2008.

[22] Hires, S, D. Tabaczynski, R, J. Noval, J, M. The Prediction of Ignition Delay and

Combustion Intervals for a Homogeneous Charge, Spark Ignition Engine. Ford Motor
Company. SAE Technical Paper 780232. 1978.

[23] Tabaczynski, R, J. Ferguson, C, R. Radhakrishnan, K. A Turbulent Entrainment Model

for Spark-Ignition Engine Combustion. Reciprocating Engines Research Department

120
Engineering and Research Staff, Ford Motor Company. Sloan Automotive Laboratory,
Massachusetts Institute of Technology. SAE Technical Paper 770647. 1977.

[24] Yeliana. Parametric combustion modeling for ethanol-gasoline fuelled spark ignition
engines. Dissertations, Master’s Theses and Master’s Report, Michigan Technology
University. 2010.

[25] Pipitone, E. A Comparison between Combustion Phase Indicators for Optimal Spark
Timing. Journal of Engineering for Gas Turbines and Power SEPTEMBER 2008, Vol. 130.
052808-1. 2008

[26] Ghanaati, A. Darus, M. Said, M. andwari, A. A mean value model for estimation of
laminar and turbulent flame speed in spark-ignition engine. University of Malaysia Pahang.
International journal of automotive and mechanical engineering, volume 11, UMP publisher.
2015.

[27] Sodré, J. A Parametric Model for Spark Ignition Engine Turbulent Flame Speed. SAE
Technical Paper 982920. 1998.

[28] Mantilla, J. Garzón, D. Galeano, C. Combustion model for spark ignition engines
operating on gasoline-ethanol blends. Engenharia Térmica Vol. 9, Universidad Nacional de
Colombia. 2010.

[29] Lipatnikov, A, N. Chomiak, J. Modeling of Pressure and Non-Stationary Effects in Spark

Ignition Engine Combustion: A Comparison of Different Approaches. Chalmers University of
Technology - Goteborg, Sweden. SAE Technical Paper 2000-01-2034. 2000.

[30] Borg, J, M. Alkidas, A, C. Characterization of Auto-ignition in a Knocking SI Engine

Using Heat Release Analysis. Automotive Products Research Laboratory, Hitachi America
Ltd. Dept. of Mechanical Engineering, Oakland University, Michigan. SAE Technical Paper
Series. 2006-01-3341. 2006.

[31] Amiel, R. Tartakovsky, L. Effect of Flight Altitude on the Knock Tendency of SI

Reciprocating Turbocharged Engines. Technion Israel Inst. of Technology. SAE Technical
Paper 2016-32-0006. 2016.

[32] Douaud, A, M. Eyzat, P. Four-octane-number method for predicting anti-knock behaviour

of fuels and engines. SAE Technical Paper No. 780080. 1978.

[33] Blizard, N. Keck, J. Experimental and theoretical investigation of turbulent burning

[34] model for internal combustion engines. SAE Technical Paper No. 740191; 1974.

[35] Ferguson, C. Tabaczynski, J. Radhakrishnan, K. A turbulent entrainment model for spark-

ignition engine combustion. SAE Technical Paper No. 770647; 1977.

[36] Tabaczynski, R. Trinker, F. Shannon, B. Further refinement and validation of a turbulent

flame propagation model for spark-ignition engines. Combustion and Flame. 1980;39:111-21.

121
[37] Kaprielian, L. Demoulin, M., Cinnella, P. Daru, V. Multi-zone quasi-dimensional
combustion models for spark-ignition engines. SAE Technical Paper No. 2013-24-0025; 2013.

[38] Van-Zwol, F. Formula student V-twin Race Engine Lubrication system. Dissertation for
MSc in Racing Engine Design. Oxford Brookes University. Student number: 06069218. 2007.

[39] Liu, C, Q. Orzechowski, J. Theoretical and Practical Aspects of Balancing a V-8 Engine
Crankshaft. DaimlerChrysler Corporation. SAE Technical Paper 2005-01-2454. 2005.

[40] Tunestål, P. Estimation of the In-Cylinder Air/Fuel Ratio of an Internal Combustion

Engine by the Use of Pressure Sensors. Doctoral thesis. Department of Heat and Power
Engineering. Lund Institute of Technology. 2001.

[41] Malaysia, N, Z. Carden, P. Bell, D. Design and Analysis of a Lightweight Crankshaft for
a Racing Motorcycle Engine, Ricardo UK. SAE Technical Paper 2007-01-0265. 2007.

[42] Jaques, P. Heat Transfer, Tribology and design of piston assembly for V-Twin Engine.
MSc Dissertation. Oxford Brookes university school of technology September. 2007.

[43] Tabaczynski, R. Trinker, F. Shannon, B. Further refinement and validation of a turbulent

flame propagation model for spark-ignition engines. Combustion and Flame. 1980;39:111-21.

[44] Nigus, H. Kinematics and Load Formulation of Engine Crank Mechanism. Mechanics,
Materials Science & Engineering, October 2015 – ISSN 2412-5954. 2015.

[45] Borg, J, M. Alkida, A, C. Investigation of the Effects of Autoignition on the Heat Release
Histories of a Knocking SI Engine Using Wiebe Functions. Hitachi Europe GmbH. Oakland
University. SAE Technical Paper series 2008-01-1088. 2018.

[46] Potul, S. Nachnolkar, R. Bhave, S. Analysis Of Change In Intake Manifold Length And
Development Of Variable Intake System. ISSN 2277-8616 international journal of scientific
and technology research volume 3, issue 5, 2014.

[47] 2016 Formula SAE® Rules, SAE International. Available at

http://fsaeonline.com/page.aspx?pageid=e179e647-cb8c-4ab0-860c-ec69aae080a3. 2016.

[48] Stone, C, R. Etminan, Y. Review of Induction System Design and a Comparison between
Prediction and Results from a Single Cylinder Diesel Engine. Brunel University. SAE
Technical Paper Series 921727. 1992.

[49] Cox, R, J, N. Formula Student Engine Design and Development – The Design and
development of a varying geometry inlet manifold. Oxford Brookes University Dissertation,
module number U04599. Student number 04091584. 2009.

[50] Helmholtz, H, A, J, E. On The Sensations Of Tone As A Physiological Basis For The

Theory Of Music. New York: Dover Publications. Print. 1954.

[51] Matins, L, R. Guimarães, G, P. Fragassa, C. Acoustical Performance of Helmholtz

Resonators Used as Vehicular Silencers. Federal University of Ouro Preto. Mechanical
Engineering Department. Brazil. 498 ▪ VOL. 46, No 4. Available at

122
https://www.mas.bg.ac.rs/_media/istrazivanje/fme/vol46/4/8_lr_martins_et_al.pdf. PDF.
2018.

[52] Mathews, R. Gardiner, A. Increasing the compression pressure in an engine by using a

long intake pipe. National Advisory Committee for Aeronautics, No.180. 1924.

[53] Capetti, A. Effect of intake pipe in the volumetric efficiency of an internal combustion
engine. National Advisory Committee for Aeronautics, No.501. 1927.

[54] Dennison, E, S. Inertia supercharging of engine cylinders, Transactions of the American

Society of Mechanical Engineers Oil Gas and Power 55 (1933) 53–64. 1933.

[55] Mariucci, V, E. An Experimental and Computational investigation of the Effect of Primary

Intake Runner Geometry on the Performance of a Single Cylinder Engine. Thesis for a master
of science degree from Ohio state University. 2006.

[56] Prosser, T, G. Induction Ramming a Motored High-Speed Four-Stroke Reciprocating

Engine—Influence of Inlet Port Pressure Waves on Volumetric Efficiency. IMechE Volume
188 pp. 577-584. 1974.

[57] Engelman, H, W. The Tuned Manifold: supercharging without a blower, ASME paper 53-
DGP-4, 1953.

[58] Winterbone, D, E. Pearson, R, J. Design Techniques for engine manifolds. Wave Action
Methods for IC Engines. Published professional Engineering Publishing Limited, London.
Printed by J W Arrowsmith Limited. UK. 1999.

[59] Winterbone, D, E. and Yoshitomi, M. The Accuracy of Calculating Wave Action in

Engine Intake Manifolds. International Congress and Exposition Detroit, Michigan February
26 - March 2,1990. Depart of Mechanical Engineering UMIST, Manchester, England and
Komatsu Limited Oyarna-shi, Japan. SAE paper series 900677. 1990.

[60] Yoshitomi, M. Investigation of wave action inside intake manifold of a reciprocating

internal combustion engine. MSc thesis. Depart of Mechanical Engineering UMIST,
Manchester, England. 1989.

[61] Pearson, R, J. A linear model for synthesis of intake manifolds. MSc thesis. Depart of
Mechanical Engineering UMIST, Manchester, England. 1988.

[62] Winterbone, D, E. Pearson, R, J. Theory of engine manifold design. Wave Action Methods
for IC Engines. Published professional Engineering Publishing Limited, London. Printed by J
W Arrowsmith Limited. UK. 2000.

[63] Broome, D. Induction Ram. Part I, II, II. Automobile Engineer. April, pp. 130-133, May,
pp. 180-184; June, pp. 262-267. 1969.

[64] Bujak, J, Z. Contribution to discussion of Kastner. Proceedings of the Institution of

Mechanical Engineers 153 June 1945. Pp 215 – 217. 1945.

123
[65] Kastner, L, J. Induction Ramming Effects in Single-Cylinder Four-Stroke Engines.
Proceedings of the Institution of Mechanical Engineers June 1945 153: 206-220, 1945.

[66] Benajes, J. and Reyes, E. Galindo, J. Peidro, J. Predesign Model for Intake Manifolds in
Internal Combustion Engines. Universidad de La Coruña and Universidad Politécnica de
Valencia. SAE Technical Paper Series 970055. 1997.

[67] Peidro, J, L. Predesign of intake manifolds in internal combustion engines. PhD thesis.
Universidad Politecnica de Valencia. 1990.

[68] Shami, C, J. Air Induction Design for Restricted Race Engines. Master of Science Thesis.
Department of Mechanical Engineering and Materials Science. Duke University. 2014.

[69] Singhal, A. Parveen, M. Air Flow Optimization via a Venturi Type Air Restrictor.
Proceedings of the World Congress on Engineering 2013 Vol III. London. 2013.

[70] Rickwood, R. Conversion of a Chevrolet LS7 engine for LMP1 use: Inlet, exhaust and
noise attenuation. Oxford Brookes University. Racing Engine Design MSc Dissertation.
Student Number 07076108. 2008.

[71] Vaughan, V. Delagrammatikas, G, J. Varibale runner length intake manifould design: an

interim progress report. Cooper union. SAE Technical Paper Series 2010–10–1112, 2010.

[72] Mackey, D, O. Crandall, J, G. Chatfield, G, F. Ashe, M, C. Optimization of Exhaust-Pipe

Tuning on a 4-Stroke Engine Using Simulation. SAE 2002 World Congress. Detroit, Michigan.
SAE Technical Paper 2002-01-002. 2002.

[73] Westin, F. and Ångström, H-E. Simulation of a Turbocharged SI-Engine with Two
Software and Comparison with Measured Data. The Royal Institute of Technology (KTH),
Stockholm. Reprinted From: Spark Ignition and Compression Ignition Engines Modelling
2003. SAE Technical Paper 2003-01-3124. 2003.

[74] Tartakovsky, L. Baibikov, V. Gutman, M. Veinblat, M. Simulation of Wankel Engine

Performance Using Commercial Software for Piston Engines. Technion Israel Inst. of
Technology. Reif, J. Elbit Systems Ltd. SAE Technical Paper 2012-32-0098. 2012.

[75] Dr Hasse, C, F. Dr Wensing, M, E. Combustion modeling for virtual SI engine calibration

with the help of 0D/3D methods. Faculty of Mechanical Engineering, Process Engineering and
Energy Engineering. Technical University Bergakademie Freiberg. Von der Fakultät für
Maschinenbau,Verfahrens- und Energietechnik der Technischen Universität Bergakademie
Freiberg genehmigte. zur Erlangung des akademischen Grades Doktor-Ingenieur (Dr.-Ing.)
vorgelegt. Von Diplom-Ingenieur Sebastian Grasreiner aus Erfurt. 2012.

124
[76] Ukidave, S, S. Development of a technique for achieving an optimum BSFC for LAF
engine. Michigan Technological University. Dissertations, Master's Theses and Master's
Reports. 2011.

[77] Yeliana. Parametric combustion modeling for ethanol-gasoline fuelled spark ignition
engines. Dissertations, Master's Theses and Master's Reports. Michigan Technological
University. 2010.

[78] Hatlevold, E, S. Assessment of engine performance and exhaust emission at changing

operating conditions and under fault conditions. Department of Marine Technology,
Norwegian University of Science and Technology. MSc Thesis. 2010-02-24. 2010.

[79] Meyer, J. INNOVATIONS IN REPRESENTATION AND CALIBRATION OF

RESIDUAL GAS FRACTION AND VOLUMETRIC EFFICIENCY IN A SPARK IGNITED,
INTERNAL COMBUSTION ENGINE. Presented in Partial Fulfillment of the Requirements
for the Degree Master of Science. Ohio State University. 2008.

[80] Kumar, M. VENKATESHMOHAN, V. Predictive Diesel Combustion Using DI-Pulse in

GT-Power. Department of Applied Mechanics, Division of Combustion, CHALMERS
UNIVERSITY OF TECHNOLOGY, Göteborg, Sweden. 2015

[81] Kakaee, A, H. Mafi, sh. Improving Simulation Accuracy of a Downsized Turbocharged

SI Engine by Developing a Predictive Combustion Model in 1D Simulation Software.
Automotive Engineering Faculty, Iran University of Science and Technology. International
Journal of Automotive Engineering. 2017

[82] Cylinder pressure traces and heat release rates of typical super-knock, conventional knock,
and normal combustion. Jpeg. Available at https://www.researchgate.net/figure/Cylinder-
pressure-traces-and-heat-release-rates-of-typical-super-knock-conventional_fig5_273919738.
2018

[83] Ukidave, S, S. Development of a technique for achieving an optimum BSFC for LAF
engine. Michigan Technological University. Dissertations, Master's Theses and Master's
Reports. 2011.

[84] Kutaeba, J, M. Yousef, S, H. Osama, H. Effect of valve lift at different IVO ,IVC and
OVERLAP angles on SI Engine performance. The 7th Jordanian International Mechanical
Engineering Conference. 2010

[85] Lumley, J. Engines, an introduction. Cambridge UK, Cambridge universities print. 1999.

[86] Badhum, T. The development of a VVT solution for the Ricardo E6 and a study of VVT
Strategy. Oxford brooks university final year project interim report. Oxford brooks university.
Student Number 11064990. 2016.

[87] Dr Samuel, S. Combustion in SI engines, U04584 Automotive Engines module, Oxford

Brookes University, 2015.

125
[88] Kim, J.Bae, C. An investigation on the effects of late intake valve closing and exhaust gas
recirculation in a single-cylinder research diesel engine in the low-load condition. Proc IMechE
Part D: J Automobile Engineering 2016, Vol. 230(6) 771–787_ IMechE 2015.

[89] The Impact of valve events upon engine performance and emissions. VALVE
EVENT.doc. Mechadyne International. 2006. Available at http://www.mechadyne-
int.com/vva-reference/papers/the-impact-of-variable-valve-actuation-on-engine-performance-
and-emissions.pdf

[90] Longfei Chen1, Richard Stone2 and Dave Richardson3 Effect of the valve timing and the
coolant temperature on particulate emissions from a gasoline directinjection engine fuelled
with gasoline and with a gasoline–ethanol blend. Proc IMechE Part D: J Automobile
Engineering 226(10) 1419–1430_ IMechE 2012.

[91] Arsie, I. Di Leo, R. Pianese, C. De Cesare, P. Air-Fuel Ratio and Trapped Mass Estimation
in Diesel Engines Using In-Cylinder Pressure. Università di Salerno. Magneti Marelli
Powertrain. SAE Technical Paper 2017-01-0593. 2017.

[92] Dr Sobachkin, A. Dr Dumnov, G. Numerical Basis of CAD-Enabled CFD. White Pape.

SolidWorks. Dassault Systems, France. Mentor Graphics Corporation Analysis Division,
Russia. PDF. 2014.

[93] Keskinen, J-P. Vuorinen, V. Kaario, O. Larmi, M. Large Eddy Simulation of the Intake
Flow in a Realistic Single Cylinder Configuration. Aalto University. SAE Technical Paper
2012-10-0137. 2012.

[94] Basara, B. Poredos, A. Gorensek, P. Scale-Resolving Simulations of the Flow in Intake

Port Geometries. AVL List GmbH. SAE Technical Paper 2016-01-0589. 2016.

[95] Chit, O, J. Omar, A, A. Asrar, A. Reynolds Averaged Navier-Stokes Flow Computation

of RAE2822 Airfoil Using Gas-kinetic BGK Scheme. Proceedings of the International
MultiConference of Engineers and Computer Scientists 2009 Vol II IMECS 2009, March 18 -
20, 2009, Hong Kong. ISBN: 978-988-17012-7-5. PDF. 2009.

[96] Cheng, S-W. Dey, T, K. Shewchuk, J, R. Delaunay Mesh Generation. International

Standard Book Number: 978-1-58488-730-0 Hardback. C7303. 2012.

[97] Fan, Q, Y. Kuba, M. Nakanishi, J. Coupled Analysis of Thermal Flow and Thermal Stress
of an Engine Exhaust Manifold. Software CRADLE Co., Ltd. SAE Technical Paper 2004-01-
1345. 2004.

[98] Hellsvik, R. Transient Simulation of Ventilation Rate and Moisture load for Cold Attic
Constructions - A CFD Analysis. Master's Thesis in Applied Mechanics. Department of Civil
and Environmental Engineering. CHALMERS UNIVERSITY OF TECHNOLOGY,
Gothenburg, Sweden. REPORT NO. 2015:57. 2015.

[99] Andersson,B. Anderson, R. Håkansson, L. Mortensen, M. Sudiyo, R. van Wachem, B.

Computational Fluid Dynamics for Engineers. Cambridge University Press. ISBN 978-1-107-
01895-2 Hardback. 2012.

126
[100] Swant,N. Yamakawa, S. Singh, S. Shimada, K. Automatic Hex-Dominant Mesh
Generation for Complex Flow Configurations. SAE Technical Paper 2018-01-0477. 2018.

[101] Valentini, P, P.Pennestri, E. Design and simulation of a variable-timing variable-lift cam

mechanism, Department of Mechanical Engineering, University of Rome ‘Tor Vergata’,
Rome, Italy. 2009.

[102] Martienz, Isidoro. Combustion Modelling of a Gasoline Engine by the Wiebe Function.
Microsoft Word Document. Available at
http://webserver.dmt.upm.es/~isidoro/bk3/c15/Combustion.pdf. 2017.

[103] Oxford Brookes Racing private team server. Edited and used by Bradshaw, B, W, P.
15105795. Current and past CAD models designed in CATIA V5R20 and Solidworks 2016.
All Intellectual Property owned by OBR and Oxford Brookes University and all rights
reserved. 2017

[104] Bradshaw, B, W, P. 15105795. CAD models designed in CATIA V5R20 and Solidworks
2016. All Intellectual Property owned by OBR and Oxford Brookes University and all rights
reserved. 2017.

[105] Mortlock, L, T, O. Investigation of a Turbocharging system on the Brookes V-Twin

Engine by 1-D Simulation with GT-Power and Its Evaluation of Performance. Student Number:
13085589. Dissertation Module Number: U04599. 2017.

[106] Franzke, D, E. Beitrag zur Ermittlung eines Klopfkriteriums der ottomotorischen

Verbrennung und zur Vorausberechnung der Klopfgrenze. Ph.D. Thesis. Technical University
of Munich. 1981.

[107] Worret, R. Bernhardt, S. Schwarz, F. Spicher, U. Application of Different Cylinder

Pressure Based Knock Detection Methods in Spark Ignition Engines. Institut für
Kolbenmaschinen, University of Karlsruhe. Germany. SAE Technical Paper 2002-01-1668.
2002.

[108] Douaud, A, M. Eyzat, P. Four-Octane-Number Method for Predicting the Anti-Knock

Behavior of Fuels and Engines. Institut Français du Pétrole. France. SAE Technical Paper
780080. 1979.

[109] McCuiston, F, D, Jr. Lavoie, G, A. Kauffman, C, W. Validation of a Turbulent Flame

Propagation Model for a Spark Ignition Engine. Department of Aerospace Engineering,
University of Cincinnati. Fuels and Lubricants Department, Ford Motor Co. SAE Technical
Paper 770045. 1977.

Picture credits

[110] Dr Yang, C. Performance and Efficiency Analysis of Engines. U04584 Automotive

Engines lecture series, Oxford Brookes University, 2015.

[111] Dr Wang, W. J. In-Cylinder Pressure Measurement and Analysis. PowerPoint

presentation, University of Essex, 2015.

127
7. Bibliography
[1] Mindworks. Engine performance parameters, University of Idaho, 2015

[2] thecartech. Engine formulas,

http://www.thecartech.com/subjects/engine/engineformulas.htm, 2015

[3] Yeliana, C. Cooney, J. Worm, D. Michalek, J. Naber. Wiebe Function Parameter

Determination For Mass Fraction Burn Calculation in a Ethanol-Gasoline Fuelled Engine,
Michigan Technological University, Journal of KONES Powertrain and Transport. 2008.

[4] Alden, C. investigation in to downsizing engine trends, limitations encountered in the

attempt to raise efficiency. Oxford Brookes University Final year project BSc motorsport
technology, department of mechanical engineering and mathematical sciences. 2015

[5] Pulkrabek, W. Engineering Fundamentals of the internal combustion engine. University of

Wisconsin. 1997

[6] Deng, B. Fu, J. Zhang, D. Yang, J. Feng, R. Liu, J. Li, K. Liu X. The heat release analysis
of bio-butanol/gasoline blends on a high speed SI (spark ignition) engine. Written by the State
Key Laboratory of Advanced Design and Manufacturing for Vehicle Body, Hunan University,
410082 Changsha, China, Department of Industrial Technology and California State
University, Fresno, Fresno, CA 93740-8002, USA, Chongqing Longxin Engine Co., Ltd,
400060 Chongqing, China. 2013.

[7] Rassweiler, G. Withrow, L. Motion Pictures of Engine Flames Correlated with Pressure
Cards. SAE Proceedings. 1938

[8] Lackner, K. Chapter 3, the concept of viscosity. Columbia University. 2002. Available at
http:/ /www.columbia.edu/itc/ldeo/lackner/E4900/Themelis3.pdf

[9] Prof Inamdar, S. Laminar & Turbulent Flows. Watershed sciences research group.
University of Delaware. 2010

[10] Ghanaati, A. Darus, M. Said, M. andwari, A. A mean value model for estimation of
laminar and turbulent flame speed in spark-ignition engine. University of Malaysia Pahang.
International journal of automotive and mechanical engineering, volume 11, UMP publisher.
2015.

[11] Sodré, J. A Parametric Model for Spark Ignition Engine Turbulent Flame Speed. SAE
Technical Paper 982920. 1998.

[13] Mantilla, J. Garzón, D. Galeano, C. Combustion model for spark ignition engines
op2erating on gasoline-ethanol blends. Engenharia Térmica Vol. 9, Universidad Nacional de
Colombia. 2010.

[13] Lower, S. Collison and activation, the Arrhenius Law, WEB available at
http://www.chem1.com/acad/webtext/dynamics/dynamics-3.html, 2009. Modified 2015

128
[14] Dogahe, M. Estimation of mass fraction of residual gases from cylinder pressure data and
its application to modeling for SI engine. Journal of Applied Mathematics, Islamic Azad
University of Lahijan, Vol.8. 2012.

[15] Blizard, N. Keck, J. Experimental and theoretical investigation of turbulent burning model
for internal combustion engines. SAE Technical Paper No. 740191; 1974.

[16] Ferguson, C. Tabaczynski, J. Radhakrishnan, K. A turbulent entrainment model for spark-

ignition engine combustion. SAE Technical Paper No. 770647; 1977.

[17] Kaprielian, L. Demoulin, M., Cinnella, P. Daru, V. Multi-zone quasi-dimensional

combustion models for spark-ignition engines. SAE Technical Paper No. 2013-24-0025; 2013.

[18] Heywood, J. B. Internal Combustion Engine Fundamentals. McGraw Hill, Automotive

Technology Series, 1988.

[19] Dr Yang, C. Performance and Efficiency Analysis of Engines. U04584 Automotive

Engines lecture series, Oxford Brookes University, 2015.

[20] Kutaeba, J, M. Yousef, S, H. Osama, H. Effect of valve lift at different IVO ,IVC and
OVERLAP angles on SI Engine performance. The 7th Jordanian International Mechanical
Engineering Conference. 2010

[21] Lumley, J. Engines, an introduction. Cambridge UK, Cambridge universities print. 1999.

[22] Kim, J.Bae, C. An investigation on the effects of late intake valve closing and exhaust gas
recirculation in a single-cylinder research diesel engine in the low-load condition. Proc IMechE
Part D: J Automobile Engineering 2016, Vol. 230(6) 771–787_ IMechE 2015.

[23] Ferguson, R. C. and Kirkpatrick, A. T.; "Internal combustion engines Applied

thermodynamics", 2nd Edition, John Wiley and sons inc., New York, 2001.

[24] Valentini, P, P.Pennestri, E. Design and simulation of a variable-timing variable-lift cam

mechanism, Department of Mechanical Engineering, University of Rome ‘Tor Vergata’,
Rome, Italy. 2009

[25] Longfei Chen1, Richard Stone2 and Dave Richardson3 Effect of the valve timing and the
coolant temperature on particulate emissions from a gasoline directinjection engine fuelled
with gasoline and with a gasoline–ethanol blend. Proc IMechE Part D: J Automobile
Engineering 226(10) 1419–1430_ IMechE 2012

[26] Nigus, H. Kinematics and Load Formulation of Engine Crank Mechanism. Mechanics,
Materials Science & Engineering, October 2015 – ISSN 2412-5954. 2015.

[27] Jaques, P. Heat Transfer, Tribology and design of piston assembly for V-Twin Engine.
MSc Dissertation. Oxford Brookes university school of technology September. 2007.

[28] Ling, J. Tun, L, T, Y. CFD Analysis of Non-Symmetrical Intake Manifold for Formula
SAE Car. School of Aerospace, Mechanical and Mechatronic Engineering. SAE Technical
Paper Series 2006-01-1976. 2006.

129
[29] Garner, C, P. Hargrave, G K. Versteeg, H, K. and Viv J. Page. Development of a Validated
CFD Process for the Analysis of Inlet Manifold Flows with EGR. Loughborough University
and Perkins Engines Company Ltd. SAE Technical Paper Series 2002-01-0071. 2002.

[30] Cui, Y. Pan, W.Leylek, J, H. and Sommer, R, G. Jain, S, K. Cylinder-to-Cylinder Variation

of Losses inIntake Regions of ICEnginesin Intake Regions of IC Engines. Clemson University
and Caterpillar, Inc. SAE Technical Paper Series 981025. 1998.

[31] Shaw, C, T. Lee, D, J. and Richardson, S, H. Pierson, S. Modelling the Effect of Plenum-
Runner Interface Geometry on the Flow Through an Inlet System. University of Warwick and
Jaguar Engineering Centre. SAE Technical Paper Series 2000-01-0569. 2000.

[32] Shinde, P, A. Research and optimization of intake restrictor for Formula SAE car engine.
Mechanical Engineering, Smt. Kashibai Navale College of Engg, Pune. International Journal
of Scientific and Research Publications, Volume 4, Issue 4, April 2014 1 ISSN 2250-3153.
2014.

[33] Davies, P. Piston Engine Intake And Exhaust System Design. Institute of Sound and
Vibration Research, University of Southampton, England. Journal of Sound and Vibration
(1996) 190(4), 677-712. 1995.

[34] Ceviz, M, A. Intake plenum volume and its influence on the engine performance, cyclic
variability and emissions. Department of Mechanical Engineering, Faculty of Engineering,
University of Ataturk, Turkey. 2006.

[35] Hwang, I, G. Myung, C-L. Park, S. and In C-B. Yeo, G, Y. Theoretical and Experimental
Flow Analysis of Exhaust Manifolds for PZEV. Korea University and Hyundai Motor
Company. SAE Technical Paper Series 2007-01-3444. 2007.

[36] Lakshmikantha, M. Kec, M. Optimization of Exhaust Systems. Eberspaecher North

America, Inc. SAE Technical Paper Series 2002-01-0059. 2002.

[37] Gonzalez, I, L. FS08 V-twin Cylinder Head Design. Oxford Brookes University. MSC
dissertation. Student Number 06072985. 2007.

[38] Bush, P. Telford, C. and Boam, D. Bingham, J. A Design Strategy for Four Cylinder SI
Automotive Engine Exhaust Systems. Arvin Exhaust, Ltd. And National Engineering
Laboratory. SAE Technical Paper Series 2000-01-0913. 2000.

[39] Martensson, J. Flardh, O. Modeling the Effect of Variable Cam Phasing on Volumetric
Efficiency, Scavenging and Torque Generation. SAE Technical Paper Series 2010-01-1190.
2010.

[40] Shaw, C, T. Lee, D, J. Tichardson, S, H. Pierson, S. Modelling the effect of plenum-runner

interface Geometry on the flow through an inlet system. University of Warwick. Jaguar
Engineering Centre. SAE Technical Paper Series. 2000-02-0569. 2000.

130
[41] Bradshaw, B, W, P. CAD work carried out by Bradshaw on Solidworks 2016/17.
15105795. 2016/17.Ferguson, R. C. and Kirkpatrick, A. T.; "Internal combustion engines
applied thermodynamics", 2nd Edition, John Wiley and sons inc., New York, 2001.

[42] Brown, B. Combustion Data Acquisition and Analysis, Department of Aeronautical and
Automotive Engineering, Loughborough University. 2014.

[43] Alden, C. investigation in to downsizing engine trends, limitations encountered in the

attempt to raise efficiency. Oxford Brookes University Final year project BSc motorsport
technology, department of mechanical engineering and mathematical sciences. 2015

[44] Salazar, F. internal combustion engine. Department of aerospace and mechanical

engineering. University of Notre Dame. 1998.

[45] Rassweiler, G. Withrow, L. Motion Pictures of Engine Flames Correlated with Pressure
Cards. SAE Proceedings. 1938

[46] Taylor, J. MCKee, R. McCullough, G. Cunningham, G. McCartan C. Computor

Simulation and Optimisation of an Intake Camshaft for a Restricketed 600 CC Four-Stroke
Engine. Queen’s University Belfast. SAE Technical Paper Series 2006-32-0071. 2006.

[47] McClintock, S. Walkingshaw, J. McCartan, C. McCullough, G. Cunningham, G.

Camshaft Design For an Inlet-Restricted FSAE Engine. Queen’s University Belfast. SAE
Techinical Paper deries 2008-32-0073. 2008

[48] McKee, R, H. McCullough, G. Cunningham, G. Taylor, J, O. McDowell, N. Taylor, J, T.

McCullough, R. Experimental Optimisation Of Manifould and Comashaft Geometries for a
Restricted 600 CC Four Clinder Four-Stroke Engine. Queen’s University Belfast. SAE
Techincal Paper 2006-32-0070. 2006

[49] Brown, B. Combustion Data Acquisition and Analysis, Department of Aeronautical and
Automotive Engineering, Loughborough University. 2014.

[50] Rassweiler, G. Withrow, L. Motion Pictures of Engine Flames Correlated with Pressure
Cards. SAE Proceedings. 1938

[51] Alden, C. investigation in to downsizing engine trends, limitations encountered in the

attempt to raise efficiency. Oxford Brookes University Final year project BSc motorsport
technology, department of mechanical engineering and mathematical sciences. 2015

[52] Mittal, M . Zhu, G .Schock , H. Fast mass-fraction-burned calculation using the net
pressure method for real-time applications. Department of Mechanical Engineering, Michigan
State University, East Lansing, MI, USA. 2008

[53] Brown, N M. Characterisation of emissions and combustion stability of a port fuelled

spark ignition engine. PhD thesis, University of Nottingham. 2009

[54] Potul, S. Nachnolkar, R. Bhave, S. Analysis Of Change In Intake Manifold Length And
Development Of Variable Intake System. ISSN 2277-8616 international journal of scientific
and technology research volume 3, issue 5, 2014

131
[55] Winterbone, D, E. Pearson, R, J. Design Techniques for engine manifolds. Wave Action
Methods for IC Engines. Published professional Engineering Publishing Limited, London.
Printed by J W Arrowsmith Limited. UK. 1999.

[56] Cox, R, J, N. Formula Student Engine Design and Development – The Design and
development of a varying geometry inlet manifold. Oxford Brookes University Dissertation,
module number U04599. Student number 04091584. 2009.

[57] Winterbone, D, E. Pearson, R, J. Theory of engine manifold design. Wave Action Methods
for IC Engines. Published professional Engineering Publishing Limited, London. Printed by J
W Arrowsmith Limited. UK. 2000.

[58] The Impact of valve events upon engine performance and emissions. VALVE
EVENT.doc. Mechadyne International. 2006. Available at http://www.mechadyne-
int.com/vva-reference/papers/the-impact-of-variable-valve-actuation-on-engine-performance-
and-emissions.pdf

[59] Badhum, T. The development of a VVT solution for the Ricardo E6 and a study of VVT
Strategy. Oxford brooks university final year project interim report. Oxford brooks university.
Student Number 11064990. 2016.

[60] Ojapah, M M. Experimental Studies of Performance and Emissions in a 2/4-Stroke Engine

with Gasoline and Ethanol. A thesis submitted for the degree of Doctor of Philosophy. School
of Engineering and Design, Brunel University London. 2014

[61] Kutaeba, J, M. Yousef, S, H. Osama, H. Effect of valve lift at different IVO ,IVC and
OVERLAP angles on SI Engine performance. The 7th Jordanian International Mechanical
Engineering Conference. 2010

[62] Nouhov, D. Investigation of the Effect of Inlet Valve Timing on the Gas Exchange Process
in High-Speed Engines, A Doctoral Thesis, Loughborough University, Institute Repository.
2004.

[63] Yang, C. Investigation of combustion and performance characteristics of CAI combustion

engine with positive and negative valve overlap. A thesis submitted for the degree of Doctor
of Philosophy. School of Engineering and Design, Brunel University. 2008

[64] Huber, R. Dynamics of Variable Valve Trains and Extrapolation Methods for Time-
Stepping Schemes. Lehrstuhl für Angewandte Mechanik, Technischen Universität München.
2012

[65] Valentini, P, P.Pennestri, E. Design and simulation of a variable-timing variable-lift cam

mechanism, Department of Mechanical Engineering, University of Rome ‘Tor Vergata’,
Rome, Italy. 2009

[66] Lumley, J. Engines, an introduction. Cambridge UK, Cambridge universities print. 1999.

132
[67] Kim, J.Bae, C. An investigation on the effects of late intake valve closing and exhaust gas
recirculation in a single-cylinder research diesel engine in the low-load condition. Proc IMechE
Part D: J Automobile Engineering 2016, Vol. 230(6) 771–787_ IMechE 2015.

[68] Ferguson, R. C. and Kirkpatrick, A. T.; "Internal combustion engines Applied

thermodynamics", 2nd Edition, John Wiley and sons inc., New York, 2001.

[69] Longfei Chen1, Richard Stone2 and Dave Richardson3 Effect of the valve timing and the
coolant temperature on particulate emissions from a gasoline directinjection engine fuelled
with gasoline and with a gasoline–ethanol blend. Proc IMechE Part D: J Automobile
Engineering 226(10) 1419–1430_ IMechE 2012

[70] Pulkrabek, W W. Engineering Fundamentals of the Internal Combustion Engine. Second

edition. Pearson education limited. 2014

[71] Gatowski, J, A. Balles, E, N. Chun, K, M. Nelson, F, E. Ekchian, J, A. Heywood, J,B.

Heat Release Analysis of Engine Pressure Data.Sloan Automotive Laboratory, Massachusetts
Institue of Technology. Fules and Lubricants Meeting & exposition Baltimore, Maryland. SAE
Technical Paper Series 841359. 1984.

133
Appendix
Model development

Figure 93: Ambient Temperature setup.

Setting up ambient conditions of 1 bar and 300 deg K

Figure 94: Initial fluid state.

Wiebe model with numbers 5 25 and 2 got from the combustion model

134
 Figure 95: Wiebe model parameters for the GT-Power model

A turbulent flame propagation model, or chemical Kinetic model, can be set up within GT-
Power but it is a highly complicated as a spark ignition map needs to be created and then
optimised for each case run, for example if a sweep is performed for the runner length then an
optimal spark map for each case needs to be created. Also the in-cylinder geometry has to be
implemented accurately within the model to allow the flame propagation model to model the
combustion phasing and flame propagation accurately.

Figure 96: Combustion model settings.

.stl files used to implement piston and cylinder head geometry into GT-Power model to set up
an accurate representation.

135
 Figure 97: Combustion chamber geometry .stl files.

Figure 98: Combustion chamber geometry input graph

The friction model

Figure 99: Engine Crank Train model.

136
And implementation of the friction model

.
Figure 100: Engine Friction model.

Engine specs and parameters/ geometry

137
 Figure 101: Engine internals’ parameters.

Inertia model Engine Effective Rotating Inertia

Figure 102: Rotating inertia of the internals.

138
Optimiser

GT-Power optimizer sweep iteration final results

Figure 103: Max Torque optimiser sweep.

Figure 104: Exhaust length optimiser sweep

139
 Figure 105: intake runner length optimiser sweep.

Figure 106: Exhaust valve timing optimiser sweep.

140
Figure 107: Intake Valve timing optimiser sweep

Figure 108: Plenum volume optimiser sweep

141
Flame Propagation

It was deemed that for this project it would be more appropriate to carry out combustion
analysis, using combustion phasing, rather than flame propagation analysis. However, the
principals of flame propagation were researched and understood, and this knowledge was
applied to combustion model as well as the combustion phasing.

Flame speed assessment

Flame velocity is one of the main ways of assessing flame propagation within a SI engine.
When the charge inside the cylinder is ignited, with the spark, a flame kernel is formed, from
there a kernel grows into a laminar flame, the expanding laminar flame forms a turbulent flame
boundary as the remaining charge burns, this turbulent flame boundary transitions the laminar
flame into a turbulent flame. Towards the end of combustion the flame can be considered
turbulent, [18].

A flame speed model for both turbulent and laminar flame speeds needs to incorporate all the
fundamentals of the SI engines combustion variables. Although the assumption, the flame
structure model doesn’t employ the flame development angle, must be made. The flame
development angle is the total crank angle between the spark ignition and the development of
the flame. [18].

The equation to explain the relationship between both laminar and turbulent flames is [18,19]–

𝜙𝜙2 𝑏𝑏
𝑆𝑆𝑇𝑇 𝑅𝑅𝑐𝑐 𝑎𝑎 𝜌𝜌𝑢𝑢 𝜙𝜙1 𝑝𝑝 𝑆𝑆𝑝𝑝̅ 𝑅𝑅𝑓𝑓
= 1+ � � × � � × � � × �0.5 × � × �1 − 𝑒𝑒𝑒𝑒𝑒𝑒 �− ��
𝑆𝑆𝐿𝐿 8 𝜌𝜌𝑏𝑏 𝑝𝑝𝑖𝑖𝑖𝑖 𝑆𝑆𝐿𝐿 𝑐𝑐 × 𝐵𝐵

Where ST and SL are Turbulent and laminar flame speed respectively. Rc is the compression
ratio, ρu and ρb are the density of the burnet and unburnt charge. p and pig are the pressure and
he pressure at ignition. 𝑆𝑆𝑝𝑝̅ is the MPS, Rf is the flame radius, B is the cylinder Bore in meters.
With a, b and c are constant while Ф1 and Ф2 are found using the equations, [18, 19] –

𝜙𝜙1 = 𝑑𝑑 + 𝑒𝑒 × (𝜙𝜙 − 1)

𝜙𝜙2 = 𝑓𝑓 + 𝑔𝑔 × (𝜙𝜙 − 1)

With d, e, f and g are adjusted using the fuel experiments procedure, [18, 19]. Ghanaati, Darus,
Said and Andwari give the values of a = 1.1, b = 0.7, c = 1.7, d = 0.611, e = 1.4, f = 1.152 and
g = -1.5, these figures where determined to allow for the calculated and measured peak pressure
had the same value and occurred at the same crank angle, [18]

The engine running and cylinder geometry parameters can be simplified from the equation for
𝑆𝑆𝑇𝑇
above into the coefficient C, [18, 19].
𝑆𝑆
𝐿𝐿

1
𝑆𝑆𝑇𝑇 𝜌𝜌𝑢𝑢 2 𝑢𝑢ˈ 𝑏𝑏 𝑅𝑅𝑓𝑓
= 1 + 𝐶𝐶 × � � × � � × �1 − 𝑒𝑒𝑒𝑒𝑒𝑒 �− ��
𝑆𝑆𝐿𝐿 𝜌𝜌𝑏𝑏 𝑆𝑆𝐿𝐿 𝑟𝑟𝑐𝑐

Laminar modelling

142
A fundamental property of the fuel/air mixture in a combustion engine is the burning velocity
normal or relative to the flame front during combustion, better known as laminar burning
velocity. [20, 21]

𝑇𝑇𝑢𝑢 𝛼𝛼 𝑝𝑝 𝛽𝛽
𝑆𝑆𝐿𝐿 = 𝑆𝑆𝐿𝐿,𝑜𝑜 � � × � �
𝑇𝑇𝑜𝑜 𝑝𝑝𝑜𝑜

Where T0 and p0 are reference temperature, 298 Kelvin, and pressures, 1 atm, respectively. Tu
is the temperature of the unburned mixture. [18, 21, 22, 27]

𝛼𝛼 = 2.18 − 0.8 × (𝜙𝜙 − 1)

𝛽𝛽 = −0.16 + 0.22 × (𝜙𝜙 − 1)

𝑆𝑆𝐿𝐿,𝑜𝑜 = 𝐵𝐵𝑚𝑚 − 𝐵𝐵𝜙𝜙 × (𝜙𝜙 − 𝜙𝜙𝑚𝑚 )2

Where Фm is the equivalence ratio at which SL,o is a maximum with a value Bm. The values of
Фm, Bm and BФ are given in the table below. SL, 0 is the reference laminar speed and used as a
flame speed correction for the variations in equivalence ratio, [18, 21, 22, 27].

Fuel Фm Bm m/s BФ m/s

Methanol 1.11 0.369 -1.405
Propane 1.08 0.342 -1.387
IsoOctane 1.13 0.263 -0.847
Gasoline 1.21 0.305 -0.549
Figure 109: fuel parameters [22, 23]

The equation for the laminar flame speed does not take into account the effects of the residual
mass fraction [18, 21, 22, 27] –

𝑇𝑇𝑢𝑢 𝛼𝛼 𝑝𝑝 𝛽𝛽
𝑆𝑆𝐿𝐿 = 𝑆𝑆𝐿𝐿,𝑜𝑜 � � × � � × �1 − 2.06 × 𝐹𝐹𝑅𝑅 0.77 �
𝑇𝑇𝑜𝑜 𝑝𝑝𝑜𝑜

Turbulence modelling
Damkohler numbers (higher the number the faster the reaction) along kolmogorov length scale
[18, 21, 22, 27] –

𝑙𝑙 𝜏𝜏𝑚𝑚 𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙ℎ𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 𝑆𝑆𝐿𝐿

𝐷𝐷𝐷𝐷 = = = ×
𝑢𝑢 × 𝜏𝜏𝑐𝑐 𝜏𝜏𝑐𝑐 𝑢𝑢′ 𝛿𝛿𝐿𝐿
𝑙𝑙
Where 𝑢𝑢′ is the mixing time, τc is the chemical reaction time, α is the thermal diffusivity, D is
the mass diffusivity, K is the thermal conductivity, Cp is the specific heat capacity, ρ is the
density, v is the kinematic viscosity, μ is the viscosity.

When the turbulence is homogeneous and isentropic then the Taylor microscale can be
represented as [18, 21, 22, 27] –

143
 � × 𝑡𝑡𝑚𝑚
𝑙𝑙𝑚𝑚 = 𝑈𝑈

The relationship between the most energetic length scale and the Kolmogorov length scale can
be represented as [18, 21, 22, 27] –
1 1
𝑙𝑙𝑘𝑘 𝜐𝜐 3 4 𝜐𝜐 3 4 −3
= � � = � � = 𝑅𝑅𝑅𝑅𝑡𝑡 4
𝑙𝑙 𝜀𝜀 × 𝑙𝑙 4 𝑢𝑢′3 × 𝑙𝑙 4

2
Where 𝐿𝐿 = �3 × 𝑢𝑢′ × 𝑙𝑙 × 𝑡𝑡, Ret is the Reynolds number when turbulent (over 4000), u’l = υt
is the viscosity of the eddy’s, u’ is the root mean square of the fluctuating velocity.

Flame front modelling

To model the turbulent flame propagation/front the controls-oriented model is used, it is based
on the quasi-dimensional turbulent flame entrainment concept developed by [26, 27, 28, 29].
The model is based on a few assumptions. The assumption can be made that the flame front
propagates through the unburned charge entraining turbulent eddies along Kolmogorov-scale
vortices. The sum of a convective component, a diffusive component, the laminar flame speed
and turbulence intensity defines the entrainment velocity. Only the leading edge and not the
whole flame area, which includes the unburned eddies, is used to define the flame front. The
turbulence intensity is an order of magnitude larger than the laminar flame speed and is defined
as the root mean square velocity of the fluctuation within the cylinder, [22, 26, 27, 28, 29].

𝑑𝑑𝑚𝑚𝑒𝑒
= 𝜌𝜌𝑢𝑢 × 𝐴𝐴𝑓𝑓 × (𝑢𝑢′ + 𝑆𝑆𝐿𝐿 )
𝑑𝑑𝑑𝑑

Where me is the mass entrained by the flame, ρu is the unburned charge density in front of the
flame, Af is the flame front area, u’ is the turbulence intensity and SL is the laminar flame speed.

The flame propagation research allowed for a better understanding of the process of
combustion while also allowing a comparison of combustion phasing and flame propagation
in terms of combustion analysis. A full analysis of combustion via aflame propagation model
would have been preferable in terms of the accuracy of the analysis but it was deemed too time
consuming and therefor combustion phasing will be sued to describe the combustion events.

144
March 2019
1
2