SHELL ECO-MARATHON: CHASSIS AND STEERING

FINAL REPORT
Group 26
June 2020
ME3 – Design Make and Test

TEAM MEMBERS
SUPERVISOR Asad Raja
Alexis Ihracska Carlos Firgau
CO-SUPERVISOR Jessica Eichel
Marc Masen Rahman Al-Shabazz

Pages: 80
EXECUTIVE SUMMARY
The aim of this project was to design, make and test the
chassis and steering system of Imperial Racing Green’s
entry to the Prototype Class of the 2020 Shell Eco-
marathon. Imperial Racing Green is an interdisciplinary
society at Imperial College London that allows students to
develop vehicles for established racing competitions. Shell
Eco-marathon is one of these competitions, in which
vehicles compete to complete a set number of laps with
the least energy consumption. The vehicle produced would Figure 1 - CAD render of final product

have to satisfy Shell Eco-marathon race regulations whilst improving upon Imperial Racing Green’s
previous entry, in particular by reducing upon its weight to improve efficiency.

The project was split into three main areas of focus: the chassis structure, the powertrain interface
and the steering system. Design development led to a spaceframe chassis made from carbon fibre
with key target dimensions of a 500 mm track width and 1550 mm wheelbase. Prototyping was used
to develop and optimise the carbon fibre tube wet lay-up joining method, and 3-point bending tests
were conducted to obtain Young’s Modulus and yield strength values for both a joint section and a
plain tube, that were used in analysis to verify the acceptability of stresses and deflections in the
chassis. The steering system’s efficiency was successfully optimised with an Ackermann geometry
configuration that was designed around target kingpin inclination and caster angles of 12° and 4°
respectively, whilst offering adjustable toe and camber angles that allowed the target of 0° for each
to be met for any driver. The final product, shown in Figure 1, was a chassis and steering system that
was lightweight, ergonomic, was successfully manufactured on-site and could fit in the shell of the
previous entry whilst allowing all powertrain components to be mounted. Its total weight of 13.85 kg
represents a 23% reduction compared to the previous entry. This exceeds the 10% target reduction.

The final testing stage was underway when the project was cut short due to the lockdown, meaning
the remaining performance testing and further regulation compliance testing that had been planned
could not been conducted. To make up for this, an alternative work package was agreed upon with
the Project Supervisor that would include spaceframe deflection hand calculations using energy
methods, a spaceframe finite element analysis and failure mode considerations and mitigations. The
sections of this report that pertain to this alternative work package have been denoted by ‘AWP’.

DMT26|| p.2
DMT26 p.2
CONTENTS
1. Introduction and Project Background ............................................................................................. 5
1.1 Competition Structure ................................................................................................................... 5
1.2 Project Aims and Objectives .......................................................................................................... 5
1.3 Shell Eco-marathon Team ............................................................................................................. 7
2. Chassis Design ............................................................................................................................... 8
2.1 Design Development ..................................................................................................................... 8
2.1.1 Wheelbase and Track Width................................................................................................... 8
2.1.2 Different Types (monocoque, spaceframe, ladder, backbone) ............................................ 11
2.2 Materials Selection ...................................................................................................................... 12
2.2.1 Materials Selection of Spaceframe Chassis .......................................................................... 12
2.2.2 Square vs Round Profile........................................................................................................ 13
2.2.3 Connection Method for Spaceframe .................................................................................... 14
2.2.4 Panel Materials Selection ..................................................................................................... 15
2.2.5 Spaceframe Joint Material Corroboration ............................................................................ 15
2.2.6 Risk Assessment ................................................................................................................... 16
2.3 Prototype Testing ........................................................................................................................ 17
2.3.1 Wet Layup Test of Joints....................................................................................................... 17
2.3.2 3-Point Bending Test: test joint vs plain tube ....................................................................... 17
2.4 Final Design Process .................................................................................................................... 19
2.4.1 Powertrain Interface ............................................................................................................ 19
2.4.2 Anthropometrics and Ergonomics ........................................................................................ 23
2.4.3 Incorporation into the Design .............................................................................................. 24
2.5 Final Chassis Design ..................................................................................................................... 25
2.6 Validation of Design..................................................................................................................... 25
2.6.1 Modes of Deflection for Automotive Chassis ....................................................................... 25
2.6.2 Spaceframe Deflections and Stresses ................................................................................... 26
2.6.3 Finite Element Analysis Corroboration of Chassis Design (AWP) .......................................... 33
2.7 Manufacturing of Spaceframe ..................................................................................................... 36
2.7.1 Chassis Assembly .................................................................................................................. 36
2.7.2 Final Layup Process............................................................................................................... 36
3. Powertrain Interfaces................................................................................................................... 38
3.1 Motor Mounting .......................................................................................................................... 38
3.2 Axle Mount Component .............................................................................................................. 39
4. Steering Design ............................................................................................................................ 42

DMT26|| p.3
DMT26 p.3
 4.1 Steering Geometry and Variable Angles ...................................................................................... 42
4.1.1 Introduction to Adjustable Steering Angles .......................................................................... 42
4.1.2 Kingpin Inclination ................................................................................................................ 43
4.1.3 Caster ................................................................................................................................... 43
4.1.4 Steering Angles Selection ..................................................................................................... 44
4.1.5 Steering Linkages Geometry ................................................................................................. 46
4.2 Interface Between Steering Input and Front Wheels .................................................................. 46
4.3 Driver Control and Actuation Method ......................................................................................... 48
4.4 Steering Design Process .............................................................................................................. 49
4.4.1 Bearing Selection and Design Considerations ...................................................................... 49
4.5 Provisional Steering Assembly Design ......................................................................................... 53
4.5.1 Wheel Spindle and ‘L’ Bracket Assemblies ........................................................................... 53
4.5.2 Brake Caliper Mounting Arm ................................................................................................ 55
4.5.3 Steering Column ................................................................................................................... 57
4.5.4 Stress Analysis of Steering Column Assembly ....................................................................... 59
4.5.5 Mounting of Steering System to Chassis .............................................................................. 64
4.6 Final Steering Assembly Design ................................................................................................... 65
4.7 Manufacturing and Assembly of Steering System ....................................................................... 66
4.7.1 Manufacturing Methods of Components ............................................................................. 66
4.7.2 Manufacturing Progress ....................................................................................................... 67
4.7.3 Mounting of Steering Assembly onto the Chassis ................................................................ 68
5. Testing and Modifications ............................................................................................................ 70
5.1 Components Inspection .............................................................................................................. 70
5.2 Ergonomics Testing ..................................................................................................................... 71
5.3 Regulation Compliance ................................................................................................................ 72
5.4 Additional Planned Testing .......................................................................................................... 73
6. Budget ......................................................................................................................................... 75
7. Discussion and Conclusions .......................................................................................................... 76
7.1 Potential Component Failure and Mitigation (AWP) ................................................................... 76
7.2 Conclusions ................................................................................................................................. 77
8. Individual Critiques ...................................................................................................................... 78
9. References................................................................................................................................... 80

DMT26|| p.4
DMT26 p.4
1. INTRODUCTION AND PROJECT BACKGROUND
Shell Eco-marathon is an international competition organised by Royal Dutch Shell aiming to develop
cleaner energy solutions and social responsibility among younger generations, providing students
with the opportunity to design, manufacture and race ultra-energy-efficient vehicles, with safety and
environmental sustainability being of primary focus. The Shell Eco-Marathon Europe event had been
scheduled to take place at Mercedes Benz World in Weybridge, Surrey from June 27 to July 3, 2020,
before it was cancelled due to unforeseen circumstances (Shell Eco-marathon, 2020).

1.1 Competition Structure

The Shell Eco-marathon event is comprised of the Prototype
Class and the Urban Concept Class. Urban Concept vehicles
resemble modern passenger cars in appearance. The main
objective for this class is to develop technologies that could be
more readily implemented in road vehicles to improve their
Figure 2 - Duke Electric Vehicles Prototype
energy efficiency. In contrast, the Prototype Class provides a entry. 14,573 mpg equivalent (Guinness
World Records, 2018)
greater degree of flexibility, focusing less on the immediate
application of technology but more on encouraging innovative solutions that result in some of the
most energy efficient vehicles of any class in the world, as shown in Figure 2.

1.2 Project Aims and Objectives

Imperial Racing Green’s previous entry to the Shell Eco-marathon Prototype Class had not met all the
regulations specified by the competition organisers. The principal aim for this project was therefore to
develop a vehicle, in conjunction with DMT Group 27, that would pass all the regulation criteria and
compete under the Prototype Class at the Shell Eco-marathon Europe event. This team designed,
manufactured and tested the vehicle’s chassis and steering system while DMT Group 27 developed its
powertrain. To ensure that this new vehicle would pass the extensive scrutineering process, the Shell
Eco-marathon official rules (Simmons, 2020) were carefully considered. As well as this, three general
project objectives (Table 1) were defined to guide the Product Design Specification (Table 2).

Table 1 - project objectives

OBJECTIVES
1. Design a chassis that will comfortably fit an average female driver, provide
mounting points for all components and be mounted to the existing aerodynamic
shell with minimal modifications.
2. Design a steering system which maximises efficiency and provides an adequate level
of feedback to the driver, allowing for adjustments to be made to the front wheel
alignment.
3. Implement a powertrain interface on the developed chassis that will be compatible
with the developments of Powertrain DMT Group 27.

DMT26|| p.5
DMT26 p.5
Table 2 - Product Design Specification
Aspect Statement / Criteria Verification
Method
Design Constraints
Dimensions
Vehicle track width At least 500 mm, between outermost contact Measurement
points of tyres and the ground (Article 39 b))
Ratio of height to track width Less than 1.25 (Article 39 c)) Measurement
Vehicle wheelbase At least 1000 mm (Article 39 d)) Measurement
Vehicle width Less than 1300 mm (Article 39 e)) Measurement
Vehicle length Less than 3500 mm (Article 39 f)) Measurement
Frame size Should fit a female driver with body Demonstrated
dimensions greater than 50% of the average in CAD
population
Vehicle body
Chassis weight At least 10% lighter than previous chassis Measurement
Vehicle deflection Maximum allowed deflection should not affect Testing
vehicle operation
Vehicle serviceability Internal components should be readily Inspection
accessible and shell easily removable
Rear compartment Easily accessible for inspection of powertrain Inspection
and to reset joulemeter
Bulkhead Upright rigid partition must separate Design review
propulsion system from driver compartment
Wheels located inside the body must also be
separated by a bulkhead.
Driver compartment Driver must be able to escape in less than 10 Testing
seconds (Article 30 a))
Steering and Brakes
Wheels Must not come into contact with the body of Inspection
the vehicle at any steering angle and testing
Turning radius Must be front wheel steered and have a Testing
minimum turning radius of 8 m or less (Article
42 a), b))
Steering set-up Wheel angle (toe and camber) must be Testing
adjustable to +/- 3° from favoured set up, with
an accuracy of 0.5°.
Brakes Must provide appropriate mounting for brake Design review
callipers
Pedal box Must provide appropriate mounting for pedal Design review
box
Safety
Chassis solidity In the event of rollover or collision the chassis Design review
will safely protect the driver’s body (Article 26
a))
Crumple zone Minimum crumple zone of 100 mm between Design review
the front of the vehicle and the driver’s feet
(Article 25 a) ii.)
Visibility Driver must have a direct arc of visibility ahead, Testing
to 90° on each side (Article 28 a)) and +/- 15°
vertically.
Mirrors Provide mounting points for two rear-view Design review
mirrors, one on each side of the vehicle
Safety belts Provide at least five mounting points for safety Testing
harness on the vehicle’s main structure that
can each withstand 200 N (Article 29 a), d))

DMT26|| p.6
DMT26 p.6
 Emergency extraction Cabin must be easily accessible and opened Inspection
from the inside and outside (Article 30 c))
Propulsion system isolation Bulkhead must provide effective isolation of Testing
battery/motor compartment (Article 27 a))
Stress safety factor Minimum safety factor of 1.2 Design review
Manufacturing Constraints
Time constraint Manufacturing must be completed by week Gantt chart
commencing 2nd March progress
analysis
Budget constraint Upper bound for project budget is £5000 Design review
Suppliers and manufacturers Use internal or approved external suppliers Design review
where possible
Materials Ensure that components can be mostly Design review
manufactured in house
Ensure correct risk assessments and
precautions are taken for toxic materials
Performance
Service life Vehicle must withstand at least 5000 km of Design review
driving distance before major maintenance

1.3 Shell Eco-marathon Team

Individual roles and responsibilities were assigned to each team member according to their previous
experiences as well as areas of interest. These are summarised in Table 3. The successful progression
of this project was also supported by members of staff and other members of the Shell Eco-marathon
team at Imperial Racing Green, summarised in Table 4.
Table 3 - team members and roles
Team Member Roles Responsibilities
Carlos Firgau Team Lead Vehicle Dynamics Analysis,
Analysis Lead Conceptual Design
Conceptualising Lead
Asad Raja Report Manager Report Writing, Time
Team Secretary Management, Stress Analysis
Prototyping and Testing Lead
Jessica Eichel Engineering Drawing Lead Engineering drawings,
Budgeting Lead Sketching, Budgeting
Embodiment Lead
Rahman Al-Shabazz CAD Lead CAD modelling,
Manufacturing Manager Manufacturing, Stress Analysis

Table 4 - external Shell Eco-marathon team

Name Roles Contact
Dr Alexis Ihracska Project Supervisor [email protected]
Dr Marc Masen Project Co-Supervisor [email protected]
Sampson Tam SEM Team Principal [email protected]
Muntadhar Mahmoud SEM Chief Mechanical Engineer [email protected]
DMT Group 27 Team Lead
Tom Mrazek SEM Chief Aerodynamics [email protected]
Engineer

DMT26|| p.7
DMT26 p.7
2. CHASSIS DESIGN
The term chassis describes the frame of a vehicle. It is responsible for providing secure locations for
and support to all vehicle components; strength and stiffness to resist all vehicle loads; and for
providing a structure that both accommodates and protects the driver (Al-Shabazz et al., 2020b). The
key design improvements in reflection of the previous entry’s chassis were to increase maximum
allowable driver height to 5’ 5” and to increase the ride height to 150 mm for better aerodynamic
performance as set by project leads. Additionally, the chassis would have to meet the requirements
outlined in Table 2, including the goal to achieve a 10% reduction in weight on comparison to the
previous entry.

2.1 Design Development

2.1.1 Wheelbase and Track Width
According to the official rules of the Shell Eco-Marathon competition (Simmons, 2019) the track width
of the Prototype Class car is defined as “the distance between the midpoints where the tyres of the
outermost wheels touch the ground”. While there is no explicit definition for the wheelbase, the
standard definition refers to the distance between the front and rear axles of the car.

The previous version of the Shell Eco-Marathon car was designed for a 4’ 8” tall driver (1.42 m tall)
and was measured to have a wheelbase of 1450 mm and a track width of 500 mm. These
measurements had to be revised to meet the driver height objectives of this project.

2.1.1.1 Wheelbase Selection

Previous results (Santing et al., 2007) indicate that the effect of wheelbase on the tyre drag value
when cornering is negligible, although a longer wheelbase means that an increased number of
components and materials are required to build the car which results in increased mass and friction
drag. However, using a shorter wheelbase can also produce negative results as it may induce flow
separation on the vehicle’s shell. After consulting with the project supervisor, it was decided that the
best method for selecting an appropriate wheelbase was to choose a value that was as small as
possible, to decrease total mass, while still being able to comfortably and safely house all the
components of the car as well as the driver. Comprehensive computational validation and testing of
prototypes in a wind tunnel were not conducted as these were outside the scope of this project.

Lengths were recorded for different sections of the vehicle from the previous chassis from which the
total approximate wheelbase was determined. To minimise the amount of material used in the
steering system, particularly in the steering column mechanism, the driver was positioned so that
their hands are as close as possible to the front wheels, without the legs extending too far forward.
Different configurations were discussed until the final arrangement was selected shown in Figure 3.

DMT26|| p.8
DMT26 p.8
 Figure 3 - length dimensions between key points

Accommodating a 5’ 5” driver required the wheelbase of the vehicle to be increased slightly to 1550
mm. Even though the increase in driver height was significant, a comparatively smaller increase in
wheelbase was achieved due to slight changes to the sitting position of the driver.

2.1.1.2 Track Width Selection

The selection of the track width for this vehicle was important as an inappropriate value could result
in rollover of the vehicle which is one of the most common types of accidents observed in fuel
economy racing. This is especially true for 3-wheeled designs due to their inherently lower stability (a
4-wheel design was discarded early on in this project as it would not fit into the teardrop shape of the
existing shell and also because it would increase the number of parts required which would lead to an
unnecessarily heavy car). Studies carried out by the ETH Zurich PAC-II team (Santing et al., 2007) also
found that aerodynamic drag increases with track width and accounts for approximately 50 % of the
total drag generated. Therefore, keeping the track width to a safe minimum became an objective.

2.1.1.2.1 Rollover Calculations

Figure 4a - axis of roll, 4b - frictional roll load component

Calculations were carried out to determine the approximate minimum track width to prevent rollover
that would be required for a vehicle operating in the Shell Eco-Marathon race conditions. Collisions

DMT26|| p.9
DMT26 p.9
for a low speed fuel efficiency race were considered unlikely and, therefore, the main mechanism for
rollover investigated was that occurring due to the vehicle attempting to turn at a very high velocity.
In this case, the axis about which the vehicle would roll is the line joining the rear wheel to one of the
front wheels, as shown in Figure 4a. The instant at which the car begins to roll is when the normal
reaction of the inside wheel, 𝑅𝑁𝐹 , is zero.

In a rotating frame of reference, the centrifugal force is equal in magnitude to the friction acting
between the tyres and the road, characterised by the coefficient of friction 𝜇. Only a component of
this which is perpendicular to the rollover axis would cause rollover, as shown in Figure 4b.

The angle 𝜃 which is the direction in which this force component acts, can also be related to the track
width, 𝑡, of the car as well as its wheelbase, 𝑏, as follows.
1
𝑡
tan 𝜃 = 2 (1)
𝑏
A balance of moments about the rollover axis at the instant the car begins to roll over, as shown in
Equation 2, allows the angle 𝜃 to also be related to the distance of the centre of mass of the car from
the rear axle when viewed from above, 𝑎 and the vertical height of the centre of mass above the
surface of the track, ℎ. The acceleration due to gravity is represented by 𝑔.
𝜇𝑚𝑔 cos 𝜃 ℎ = 𝑚𝑔𝑎 sin 𝜃 (2)
𝜇ℎ
∴ tan 𝜃 = (3)
𝑎
Equating both expressions of tan 𝜃, it was possible to obtain an expression for the minimum track
with necessary for the car to avoid rolling over.
2𝑏ℎ𝜇
𝑡= (4)
𝑎
From this steady state turning analysis, an estimated minimum track width was calculated using a
2
wheelbase, 𝑏, of 1.55 m. The height of the centre of mass, ℎ, was estimated to be approximately of
3

the radius of the wheel (250 mm) – an acceptable assumption as the majority of the driver’s body was
expected to sit below the plane of the centres of the wheels in the final design. 𝜇 was approximated
to be 0.75, a realistic estimate for the tyres available in the Pit Garage which have a lower quality of
rubber compound compared to the Michelin 45-75R16 used by ETH Zurich which had a 𝜇 of 0.8
1
(Santing et al., 2007). 𝑎 was taken to be approximately 3 of the wheelbase, measured from the front

axles – an acceptable assumption since the majority of the driver’s body was expected to be closer to
1
the front wheels (3 is proven to be an acceptable approximation in the ‘Spaceframe Deflections and

Stresses’ section, where an 𝑎 value of 0.42 was calculated). A minimum acceptable track width of 381
mm was obtained.

DMT26| |p.10
DMT26 p.10
In this analysis, the minimum trackwidth is independent of the velocity and turn radius of the car. The
underlying assumption is that the turn is taken at the maximum velocity allowed by the grip of the
tyres – any higher cornering velocity would only cause the vehicle to slide instead of rolling over.

It is evident that the minimum track width necessary is significantly below the 500 mm minimum
allowed by the rules. To avoid an unwanted drag penalty, a value of 500 mm was selected. In practice,
ℎ
this meant that a maximum ratio 𝑎
of 0.215 could be imposed on the design of the chassis to ensure

that this trackwidth would prevent rollover for the wheelbase and 𝜇 stated previously.

2.1.2 Different Types (monocoque, spaceframe, ladder, backbone)

Following from the research on loading cases during operation of the vehicle, the next step was to
decide whether a ladder, monocoque or spaceframe chassis design would be best suited for a
successful design.

2.1.2.1 Ladder
The most basic chassis structure is the ladder frame design, named after its resemblance to a
conventional ladder (Waterman, 2011). This structure is built from two members that run parallel
through the length of the vehicle, connected by multiple members perpendicular to the chassis
direction. This design dates back to horse-drawn carriages, traditionally made from wood. However,
its poor torsional stiffness made it unsuitable for this car.

2.1.2.2 Monocoque
A monocoque chassis design carries the potential for a stiffer and stronger chassis structure (Seward,
2014). It makes use of plates and shells loaded in shear to construct a closed box that acts as a
protective outer body work with a rigid internal structure that is able to carry all loads. The benefit of
using entire panels and shells is a higher polar moment of inertia in the longitudinal direction. This
reduces torsional stresses (Waterman, 2011) which promotes better handling and a safer design.

With a high material consumption, special material composites are often used to reduce the
compromise on weight. Carbon fibre composites have recently been introduced as a means of
enabling a lightweight, smooth yet strong structure. The manufacturing process for carbon fibre
composites is complex. The first step is to create a full-size mould of the chassis using wood or a resin.
With this mould made, there are two approaches to completing the monocoque shell (Seward, 2014).

1. Wet lay-up – applying alternating layers of carbon fibre and resin to the mould and allowing
to cure in a vacuum bag before detaching from the mould.

DMT26| |p.11
DMT26 p.11
 2. Pre-Impregnated (Pre-preg) – applying a premade carbon fibre-resin reinforcement to the
mould and allowing to cure in a vacuum bag inside an autoclave before detaching from the
mould. Generally, this process generates the lightest and strongest monocoque designs.

2.1.2.3 Spaceframe
The spaceframe design involves creating a 3D structure from tubes. By ensuring all members are
joined at nodes with the use of diagonal members to form triangles, stresses can be reduced by at
least one order of magnitude (Waterman, 2011). Any loads on the structure should then be applied
primarily at nodes. This is to ensure that all members in the structure are loaded in pure tension or
compression. Depending on the material, joining techniques that may be applied between members
include welding, carbon fibre wet lay-ups and internal or external lugs. The driver’s seat in a
spaceframe structure is comprised of floor panels attached between the spaceframe members.

2.1.2.4 Selection
With a low torsional stiffness, the ladder approach was deemed unviable for the chassis. The
monocoque design was ruled out immediately as no vacuum bag or autoclave sufficiently large to
enclose a full chassis structure were available on-site. Both, the wet lay-up and the pre-preg process
would have had to be outsourced – a costly approach. Therefore, it was concluded that a spaceframe
design was the most suitable option. A dual support member structure was preferable over a
backbone structure as the loads experienced on the chassis would be distributed across two members
rather than one (Al-Shabazz et al, 2020b). As well as this, due to the material selection of carbon fibre
made further into the development of the chassis to minimise the overall weight, the backbone
structure was not appropriate as it would require all members to be joined to the single backbone
structure – a configuration that favours welding. Carbon fibre cannot be welded and so would have
raised difficulties while manufacturing.

2.2 Materials Selection

The Shell Eco-Marathon is a fuel efficiency race where a high strength to weight ratio is desirable in
order to travel furthest on the fuel supplied whilst withstanding the load condition. Other desirable as
well as regulated criteria are the chassis’ ability to withstand a 700N force at its roll bar (bulkhead),
designed to protect the driver in the event of rollover, and for the chassis to experience minimum
deflection.

2.2.1 Materials Selection of Spaceframe Chassis

The properties and potential of metals and carbon fibre for building a spaceframe are given in Table 5.

DMT26| |p.12
DMT26 p.12
Table 5 – Potential of various materials for the spaceframe
Material Youngs Yield strength Density Comments
modulus (GPa) (MPa) (g/cm3)
SAE ASU 200 470 7.85 Easily machined using all conventional techniques
4340 Steel Not available as tube
Too heavy
IOM Steel 200 496 8 £2.50 per foot
Tube Too heavy
4130 205 517 7.85 £8.50 per foot
Chrome Requires pre and post welding treatment
moly Too heavy and complex to manufacture
Docol Tube 200 690 7-8 Easily welded
R8 Higher specification than required
Expensive and heavy
Carbon 90 Varies with 1.6 Connect using joining techniques in Table 7
Tubes tube £20 per metre
Circular or square tubes
Aluminium 72 35 2.7 £13.47 per m
Can be easily welded and machined
Various shaped tubes

Carbon fibre and aluminium were evidently the most suitable choices given their low density, ease of
joining by welding or numerous carbon fibre joining techniques, and availability to purchase. To help
inform the design decision, tube shapes of both materials were analysed next.

2.2.2 Square vs Round Profile

One decision to consider was the use of square vs round tubes. A simple calculation was used to
compare the mass per unit length of the two cross sections shown in Figure 5 (note the discrepancy in
thickness analysed reflecting the purchasing options available). At the time of analysis, carbon fibre
and aluminium were the two most favourable materials and so were further analysed.

Figure 5 - analysed cross-sections

The results (Table 6), showed that a circular cross-section would be highly beneficial from the weight
reduction perspective. The carbon fibre tubes also displayed a 40% weight reduction compared to the
aluminium equivalents, increasing efficiency.

Table 6 – Comparison of square and circular tubes

Square Circular
Carbon Fibre Aluminium Carbon Fibre Aluminium
Mass (kg/m) 1.776 2.997 0.955 1.612

DMT26| |p.13
DMT26 p.13
Aluminium would require cutting and spot welding to create a spaceframe, while carbon fibre would
require one of the techniques in Table 8. Due to health concerns within Imperial College London,
welding would be problematic. Therefore, it was decided to use round carbon fibre tubes for the
spaceframe. This was also more cost effective, with round tubes costing £8 less per metre than their
square tube equivalent (Easy Composites, 2019).

2.2.3 Connection Method for Spaceframe

The two fundamental joining options for the carbon fibre spaceframe tubes were as shown in Table 7.
Table 7 - Joining methods for carbon fibre tubes
Joining Method Process Comments
Carbon fibre wet layup Hold together with a jig and wrap Jig is cost and labour
around held together tubes joints with carbon fibre pre-preg or intensive.
carbon fibre fabric with an applied Difficult to ensure angles
epoxy resin mixture. correct for more complex
spaceframe geometry.
Lugged tube connectors Join using lugs made from carbon, Angular arrangement of
steel or titanium. tubes limited by connector
available.

A weight analysis was carried out on a basic chassis design (Figure 6) to explore the effect of both
connection types. The analysis was based on steel connector lugs that were available as the cost to
externally source customised carbon, aluminium or titanium lugs would exceed the budget.

Figure 6 - weight analysis spaceframe model

The analysis revealed that using steel connectors would result in a chassis weight of approximately
14.4 kg, with over 85% of that weight coming from the weight of the steel connectors. Alternatively,
the weight of the chassis using the carbon fibre wet lay-up connection method was found to be
approximately 3.0 kg. Thus, the wet layup method was pursued for further design developments.

Analysis and testing of prototype joints supported the use of carbon fibre tubes with wet lay-up
joining. The joining method was developed to use 3-D printed polymer lug inserts for alignment of
members around joints instead of a jig. This allowed for a more complex spaceframe design.

DMT26| |p.14
DMT26 p.14
 2.2.4 Panel Materials Selection
The floor and back panels provide support for the weight of the driver, separate the driver from the
rear compartment and shell floor, and are used to mount the pedal box. Table 8 shows suitable panel
materials which are readily available, simple to manufacture into the required shape, have low
density and have acceptable specific strength and stiffness.
Table 8 - Sheet materials and properties considered for panels
Material Mechanical Properties Ease of Fabrication Cost Weight
Aluminium Low density for metal Very easily machinable using ~£40/sqm 11kg / sqm
Sheet all conventional tools
Carbon Lightweight, stiff Can be manually cut. Ensure ~£300/sqm 2.95kg/sqm
Aluminium Low longitudinal modulus safety precautions are
Sandwich Panels High transverse shear modulus followed for carbon fibre
Carbon Foam Lightweight, stiff Can be manually cut. Ensure ~£180/sqm 2.71kg/sqm
Core Sandwich High longitudinal modulus safety precautions are
Panels Low transverse shear modulus followed for carbon fibre

Carbon sandwich panels with a closed-cell PVC foam core were selected for the floor panels, due to
their high stiffness and lower cost and density than aluminium cored panels or aluminium sheets,
(Figure 7).

The roll bar or bulkhead was made of a 4 mm aluminium plate, to ensure it would endure a rollover
impact and a 700 N force from all directions. It is also fire retardant, which was a race requirement to
ensure safety since it would separate the powertrain compartment from the driver.

Figure 7 - floor panel and bulkhead locations

2.2.5 Spaceframe Joint Material Corroboration

Carbon fibre fabric with an applied epoxy-resin mixture was used for the wet-layup spaceframe
joining method as opposed to carbon fibre pre-peg which would require the use of an autoclave.
Carbon fibre reinforced polymers are composite materials where the carbon fibre provides the
strength and stiffness while the polymer, usually an epoxy resin matrix, transmits loads to the fibres
and protects and holds them together (Fekete, 2017). Long strands of carbon filament are grouped
together and bonded with an epoxy resin matrix to form a laminate. A laminate is strongest when
loaded in the fibre orientation, hence multidirectional weaving patterns can increase the torsional

DMT26| |p.15
DMT26 p.15
stiffness, flexural rigidity and tensile strength. Multiple laminates are layered together in a compact
structure which minimises voids, providing a very light, strong and fatigue resistant component.

Table 9 shows the Young’s modulus and thermal expansion coefficient for different composite
materials with potential in high strength and low weight applications (King, 1989).
Table 9 - Youngs modulus and thermal expansion coefficient for composite materials

Material Young’s Modulus Thermal expansion

(GNm-2) coefficient (𝜶 × 𝟏𝟎−𝟔 °𝑪−𝟏 )
Carbon fibres – along fibre 200-800 -0.1—1.2
Carbon fibres – perpendicular to fibres 10-20 7-12
Glass fibres 76 4.9
Epoxy resins 3-6 70
Polyester resins 2-4.5 100-200
Carbon fibre / epoxy unidirectional laminae – parallel to 220 -0.2
fibres
Glass fibre / polyester unidirectional laminae – parallel to 38 11
fibres

It is evident that carbon fibres bonded with an epoxy resin matrix have a much higher Young’s
modulus than glass fibre with a polyester matrix. Carbon fibres also have a much lower thermal
expansion coefficient, corroborating them as suitable for automotive use where heat from tyre
friction and the powertrain raises temperatures.

2.2.5.1 Factors affecting strength

The strength of a carbon fibre reinforced polymer component is particularly sensitive to interface
bonding. Greater interface bonding arises from stronger adhesion of the epoxy matrix with the fibre,
achieved by ensuring carbon strands are fully covered with epoxy resin to prevent voids forming.

2.2.5.2 Failure modes

Although carbon fibre epoxy composites typically have a very long fatigue life and can withstand
many cycles of stress before failure, the structures can vary and will dictate if and how a component
will fail. During loading and repeated use, stress will accumulate, the fibre matrix interface will
weaken the bonds, and any voids present will act as microcracks and grow. This can lead to failure
mechanisms such as fibres bending, buckling and fracturing or failure of epoxy matrix. The amount of
damage accumulated depends on the matrix properties and the strength of the reinforcing fibres.

2.2.6 Risk Assessment

Having a carbon fibre-based design meant that the team would be working with processes seldom
used by other DMT groups. Thorough research was conducted to ensure that the project was
completed with the necessary safety equipment and precautions. These precautions were outlined in
relevant risk assessment forms that explored the dangers of working with carbon fibre and resins.
Each team member completed a mask fitting test and briefing with a safety officer at Imperial.

DMT26| |p.16
DMT26 p.16
2.3 Prototype Testing
2.3.1 Wet Layup Test of Joints
The objective of performing multiple prototype tests was initially to test the feasibility of the wet
layup method to connect carbon fibre tubes together. It was found that by careful preparation of
surfaces to increase adhesion, and optimising the amount of fibre and epoxy used, the joints would
cure with the strongest matrix bonds. Prototyping helped the group finalise the process used.

The prototypes produced are shown in Table 10. Straight membered joints were made by clamping
the tubes in place. Angled dual member joints were also made to test the feasibility of more complex
geometries. However, clamping the tubes accurately for complex geometries was found to be
problematic. Therefore, plastic lug inserts were prototyped using the UPS Mini 3D printers in the
Imperial College Advanced hackspace and inserted before joining to align two tubes in a 90° joint.

Table 10 - Joints prototyped

Prototype Class 1: Straight Joint Prototype Class 2: Angled Joint Prototype Class 3: Angled Joint
without Lug using 3D Printed Lug Insert

Figure 8a Figure 8b Figure 8c

2.3.1.1 Improvements and Optimisation

Cutting and filing of tube profiles such that tubes sat flush at joints and covered the lugs completely
was considered. However, prototypes showed that the effect of a slightly exposed lug was minimal,
and that the strength of the joint was more dependent on the wrapping and layup around it.

Cutting strips from a large sheet of carbon fibre weave was found to be too delicate to manually
handle. Therefore, rolls of woven carbon fibre pre-cut to widths of 25 and 50mm with secured edges
were purchased, helping to prevent misalignment and fraying during layup. Since a vacuum bag could
not be used for such a large assembly to tightly cure the layup, a composite shrink tape was used
instead, which was wrapped around joints and shrunk on heating. It was important to apply the
correct amount of resin – too much and there was unwanted liquid during wrapping causing excess
slip between layers; too little and bonds did not fully cure.

2.3.2 3-Point Bending Test: test joint vs plain tube

After prototyping a straight carbon fibre tube joint as shown in Figure 8a, a 3-point bending test was
conducted using the Instron-3366 testing machine (Figure 9). A 3-point bending test was also carried
out on plain carbon fibre tubes under the same test conditions for comparison. The results are shown

DMT26| |p.17
DMT26 p.17
in Figure 10 as force-deflection graphs. The aim of this test was to calculate and compare the Young’s
Modulus and yield strength of a joint section and that of a plain tube. These material properties were
needed for stress and deflection calculations on the spaceframe.

2.3.2.1 Method
1. The 3-point bending test blocks were attached to the Instron test machine, as shown in Figure
4. The support positions were adjusted to be 200mm apart and equidistant from the load
head with a deflection limit set to 150mm.
2. The test specimen was placed on the supports,
making sure it was straight and that the centre of
the joint sat directly underneath the load head.
3. The machine loading was started, and recordings
were taken of the load and deflection values at a
rate of 10 samples per second.
4. The test was repeated for two plain carbon tubes
Figure 9 - 3-point bending on test joint
without a joint.
5. Load versus deflection graphs were plotted for each test and used to calculate Young’s
Modulus and yield strengths for both the joint section and a plain tube.

2.3.2.2 Results

Figure 10 – graphical analysis from bending tests

2.3.2.3 Interpretation of Results

According to the ISO 13586:2000 standard, linear elastic fracture mechanics can be used for a 3-point
bend test of a composite specimen if the non-linearity of the load-displacement line is less than 10%,
as described by Equation 5.
𝐹𝑚𝑎𝑥
< 1.1 (5)
𝐹5%

DMT26| |p.18
DMT26 p.18
The force-deflection graphs show that the plain carbon tubes satisfy this criterion, and therefore
linear elastic fracture mechanics was used to analyse approximately the yield strength and Young’s
Modulus (Table 11). For the wrapped test joint, the graph did not show fully linear loading. Instead, to
approximate the loading as linear, upper and lower bound loading lines were drawn. For Equation 5
to hold true, and to give the most conservative values of yield strength and Young’s modulus, the
lower bound value of 11.9 KNm-1 was used as the gradient of the test joint.

Table 11 - 3 point bending test results

Calculation Plain Carbon Tube (average of two) Test Joint
Pmax (N) Maximum value on graph 810 390

Maximum Bending 𝑃𝑚𝑎𝑥 𝐿

𝑀𝑚𝑎𝑥 =
Moment (Nm) 4 40.5 19.5
𝑃𝑚𝑎𝑥 = 𝑈𝑇𝑆
𝐿 = 0.15 𝑚
𝑀𝑦
𝜎𝑦 =
Yield Stress (MPa) 𝐼 153.6 71.8
𝜋
𝐼 = (0.014 − 0.0094 )
4
Gradient (KN/m) ∆𝐹 𝑊 385 11.9
𝐺𝑟𝑎𝑑𝑖𝑒𝑛𝑡 = =
∆𝑑 𝛿
3
𝑊𝐿
𝛿=
Youngs Modulus 48𝐸𝐼 23.86 0.734
(GPa) 𝑔𝑟𝑎𝑑𝑖𝑒𝑛𝑡 ∗ 𝐿3
𝐸=
48𝐼
2.3.2.4 Conclusions
The yield strength of the tube with the test joint was found to be half that of a plain carbon tube and
the Young’s Modulus of the test joint was also considerably lower. As the test joint was the first
prototype produced, likely causes are the failure modes and factors affecting strength mentioned
previously such as poor adhesion and insufficient epoxy matrix applied. Future joints followed a
sophisticated layup method as described in the ‘Manufacturing of Spaceframe’ section.

2.4 Final Design Process

2.4.1 Powertrain Interface
An early design decision that had to be made in order to avoid inconsistency and conflict with design
decisions made by the Powertrain DMT team was the nature of the chassis’ interface with powertrain
components. The powertrain interface would be concerned with the design of the chassis’ rear
section behind the bulkhead, positioned just behind the driver’s head. To make a start on the
conceptualisation, measurements (Figure 11) of the previous entry were taken (Table 12) to define
the approximate dimensional scale of the rear chassis section.

DMT26| |p.19
DMT26 p.19
 Table 12 - powertrain interface measurements
Measurement Value (mm)

Bulkhead Width 400

Bulkhead Height 300

Bulkhead to Axle 500

Figure 11 - powertrain interface measurements

It was clear that the width of this section would have to narrow from the bulkhead to the axle to
ensure that it could fit within the teardrop shape of the shell. The width of this rearmost part would
therefore be constrained by the rear axle length. This dimension, as well as all the other relevant
information that was obtained from the Powertrain DMT team, is summarised in Table 13. Many of
these final dimensions were derived through design iterations by the Powertrain team based on
mutually informed decisions. Early meetings with the Shell Eco-marathon Team Principal revealed
that one of the improvements that this chassis design should make upon the previous was a less
cramped rear section to allow for easier access to and assembly of the powertrain components.

Table 13 - powertrain interface dimensions

Powertrain Parameter Value

Components Battery, Motor, Motor Controller, Axle

Type of Drive Chain

Axle Dimensions 10 mm diameter, 135 mm long

Motor Dimensions 52 mm diameter, 235 mm long

Battery Dimensions 390 mm x 70 mm x 110 mm

Motor Controller Dimensions 80 mm x 80 mm x 30 mm

Motor Speed 600 rad/s

2.4.1.1 General Powertrain Interface Design

Five different configurations were conceptualised for the powertrain interface design. These are
summarised in Table 14 along with their relative merits and drawbacks.

DMT26| |p.20
DMT26 p.20
Table 14 - conceptualised powertrain interface designs
Configuration ① Two-Section ② Simple Rear ③ Panel Design ④ Axle ⑤ Aluminium
Rear Spaceframe Spaceframe Mounted to Shell Sheet
Mounting Plate

Explanation The rear The rear Brackets extend The axle is An aluminium
spaceframe spaceframe from the back of mounted directly sheet is bent
extends from the extends from the bulkhead to the shell at a around a
bulkhead in a the bulkhead and plastic hollow slotted transverse
large cube into a narrowing panels are section in the member in the
structure. triangular attached to shell’s sandwich rear
Mounting plates structure to the them. The motor panel. spaceframe.
can be attached axle. and axle are Components are Components
between Components are mounted mounted to the are mounted to
members in this mounted to the between the back of the the aluminium
section. Members back of the plastic panels bulkhead, with a sheet and the
extend from this bulkhead. and other short protruding motor is
section in a components are mount for the mounted to the
narrowing mounted to the motor. transverse
triangular shape panels. member.
to the axle.

Sketches

Pros - Lots of space for - Saves weight - Minimises - Saves weight - Lots of space
mounting of and space. number of joints and space. for mounting
components. – easier to make components.
- Reduces and less chance - No joints –
- Similar to the number of joints of misalignment. easier to make. - Can design
previous entry so – easier to make sheet to mount
mounting panels and less chance - Plastic panels everything
can be reused. of misalignment. easy to compactly
manufacture whilst using cut
and metal outs to save
brackets readily weight.
available.

Cons - Required many - May not be - Unnecessarily - May prove - Bending of the
joints which may enough space to complicated difficult to mount aluminium
introduce mount all the motor mounting the axle in the sheet may be
misalignment. components. that does not shell due to its an inaccurate
make use of its curved shape. process.
- May be too large foot mount
to fit into shell. bracket. - Cutting the
sandwich panel
- Unnecessary - May be too may be
weight and space large to fit into challenging.
contribution the shell.
- May not be
enough space to

DMT26| |p.21
DMT26 p.21
 mount all the
components.

- Flexing of shell
due to loads may
lead to chain
misalignment.

Using the Product Design Specification as a guideline, each idea was evaluated using a decision matrix
with the following four categories: space for mounting all components whilst fitting in shell, weight,
reliability of chain alignment and ease of manufacture.

Weightings were assigned to each category from 1 to 4, with 4 signifying the most important category
for the application. Each idea was then ranked from 1 to 5, with 5 being assigned to the idea that best
fulfils the requirements of each category. The decision matrix is shown in Table 15.

Table 15 - morphological analysis of powertrain interface designs

Option 1 Option 2 Option 3 Option 4 Option 5

Weighted Weighted Weighted Weighted Weighted

Criteria Weighting Score Score Score Score Score Score Score Score Score Score

Space 4 2 8 4 16 1 4 3 12 5 20

Weight 1 1 1 4 4 2 2 5 5 3 3

Reliability of
Chain
Alignment 3 3 9 4 12 2 6 1 3 5 15

Ease of
Manufacture 2 3 6 5 10 2 4 1 2 4 8

Total Total Total Total Total

1: 24 2: 42 3: 16 4: 22 5: 46

This shows that design options ② and ⑤ were the best out of those considered. These designs
were therefore integrated and developed. Specifically, a transverse member was added to the simple
rear spaceframe of concept ② and the aluminium sheet of concept ⑤ was replaced with a
mounting plate for the motor, with the battery mounted to the back of the bulkhead. The developed
concept was integrated with the rest of the spaceframe as shown in Figure 12. This sketch shows a
slotted rear axle mount component, which was considered in order to offer adjustment of the axle
position such that the chain could be tightened. In the final powertrain interface design shown in
Figure 13, this has been removed as chain tightening was instead offered by slots in the motor’s foot
mount bracket designed by the Powertrain DMT team. The final design also developed the rear
spaceframe structure further such that the rear-most spaceframe point where the axle was to be
mounted was made to be a node of the spaceframe, to allow for better stress distribution.

DMT26| |p.22
DMT26 p.22
 Figure 12 - powertrain interface developed design
Bulkhead

Slotted motor
Brake Caliper
foot mount
Battery Mount
bracket for chain
tightening Motor

Motor Mount Figure 13 - final powertrain interface design Axle Mount

Plate

2.4.2 Anthropometrics and Ergonomics

A significant proportion of road accidents are the result of driver fatigue; with the severity of injuries
also influenced by the ride position of the victims in the vehicle. Fatigue was particularly relevant in
this case, given the tight space constraints and the length of the race with an estimated driving time
of one hour. In order to reduce risks, vehicle ergonomics had to be considered with a view to promote
ride comfort, reducing fatigue, whilst considering the driver’s spatial consciousness (Kovacevic, 2010).

As outlined in the PDS, the requirement was for the vehicle to hold a 5’ 5” female driver whose
dimensions were examined. The following results were found (Kovacevic, 2010).

Table 16 – 5’ 5” female driver dimensions

Dimension Name Length (mm)
Buttock height 854
Buttock-knee length 600
Elbow rest height - standing 1014

DMT26| |p.23
DMT26 p.23
 Elbow centre of grip length 335
Eye height – sitting 741
Forearm breadth 480
Functional leg length – seated 1030
Hip breadth 350
Shoulder height - standing 1354
Height 865
Height stature 165

2.4.3 Incorporation into the Design

With the measurements found through research, a scaled design was produced for the lateral cross
section of the vehicle (Figure 14). The design was developed around the requirement of the driver’s
knees having to be in line with the front wheel axles. This was necessary to ensure the drivers legs did
not interfere with the steering system. The design also incorporated the 150 mm ride height to
enhance aerodynamic performance (set by the Aerodynamic Lead of the project) and the general
dimensional decisions described in the ‘Wheelbase and Track Width’ section.

Figure 14 -phase 1 lateral dimensions schematic

With the foundation for the design set, the initial design was iterated to integrate the powertrain
interface, accommodate the pedal box, increase the triangulation of members for better stress
distribution and to ensure that the chassis would fit in the shell as shown in Figure 15

Figure 15 - design iterations

DMT26| |p.24
DMT26 p.24
The anthropometric data of Table 16 was then reconsidered to develop the resulting design (Figure
16), adjusting the length of relevant sections so that the driver’s body parts could sit in the locations
indicated to ensure that the main loads of the driver’s weight were at nodes.

Figure 16 - anthropometric considerations

2.5 Final Chassis Design

The final chassis layout is shown in Figure 17, with numbers labelling each node. Note that this
labelling scheme is used throughout the ‘Validation of Design’ section. This final design incorporates
the considerations of ease of manufacture, strength, weight, ergonomics, shell fit and compatibility
with powertrain components.

Figure 17 - final spaceframe with nodal annotations

2.6 Validation of Design

2.6.1 Modes of Deflection for Automotive Chassis
A structurally stable chassis would have to mitigate against the typical automobile deflection modes
summarised in Table 17.

DMT26| |p.25
DMT26 p.25
Table 17 - failure modes
Mode Description
Longitudinal Torsion Results from an imbalance of forces on opposite
corners of the vehicle. It is experienced at every corner
of a track due to the lateral weight transfer that occurs
when cornering. During this loading case, the chassis
can be modelled as a torsional spring that connects the
two corners where the loads act.

Vertical Bending Results from any vertical loads, including those induced
from vertical acceleration, acting on the structure, with
the main contribution coming from the weight of the
driver. The reaction to these forces is distributed across
the axles, where contact with the floor is made through
the tires. In this loading case, the frame can be
modelled as a simply supported beam – see ‘Simply
Supported Beam Model of Chassis’.

Lateral Bending Similar to longitudinal torsion, the main cause of these

loads is due to centrifugal forces when cornering. Side
winds and the road camber can also cause lateral
bending. These lateral forces are resisted by friction
between the tyres and the road, causing bending.

Horizontal Lozenging Horizontal lozenging occurs as a result of forward and

backward forces that are applied at opposite wheels.
These forces are due to the continuous impact of the
wheels with potholes, curbs, and general vertical
variations in the road.

Generally, if the torsional stiffness and vertical bending stiffness are satisfactory, then the resistance
to lateral bending and horizontal lozenging will also be sufficient (Singh, 2010). The following sections
therefore focus on vertical bending and longitudinal torsion as well as stresses in individual members.

2.6.2 Spaceframe Deflections and Stresses

It was important to find the amount of vertical deflection in the spaceframe due to applied forces as
excessive deflection would potentially have an adverse effect on the ride height, driver ergonomics,
chain alignment and steering system angles. If the spaceframe deflection exceeded the 0.5 m

DMT26| |p.26
DMT26 p.26
clearance gap between the chassis and shell, abrasion of tubes and the shell interior may result as
they rub. When designing the spaceframe, a simply supported beam model was used to find the
deflection in either side section of the spaceframe, with the wheels representing simple supports. It
was reasoned that, should the deflections be acceptable in this oversimplified model, the deflection
of the actual chassis would certainly be acceptable as its truss structure is known to have better
deflection behaviour than a beam (Lin, 2017). This was verified by finding more accurate deflection
values by applying an energy method for a truss structure.

Given that the tubes can be modelled as being pin-jointed to each other, the only loads that will be
experienced by the majority of the tubes will be axial tension and compression. This is a benefit due
to the very strong tensile and compressive load carrying capacity of the carbon fibre tubes chosen.

2.6.2.1 Loading Condition

In its static loading condition, the chassis is subject to the weight of the driver, the weight of the pedal
box and the weight of the powertrain components. The Powertrain DMT team estimated the pedal
box and powertrain components to be 10 kg each, and the driver mass that the vehicle had been
designed for was 60 kg. The centre of mass of a person on average lies between their stomach and
chest (Grabarczyk, 2003). The centre of mass of the driver was therefore modelled as lying between
nodes 9/10 and 13/14. As shown in Figure 18, this was considered along with the approximate
locations of the pedal box and powertrain centre of masses to calculate a total centre of mass
location of 1.2 m from the front most member of the spaceframe and 0.65 m from the front wheels.
Front Wheels Node 9/10 Node 13/14 Rear Wheel
1.55 m

10g 80g 60g 10g

0.65 m
1.2 m
1.29 m
1.65 m
Driver CoM
Pedal Box CoM Overall CoM Powertrain CoM

Figure 18 - determining centre of mass

The dynamic loading conditions are more difficult to estimate as they will be determined by shock
loads from bumps in the road. In literature, this is accounted for by applying a dynamic multiplication
factor to all loads (Seward, 2014). For the simply supported beam model, this was neglected since the
purpose of this model was not to give an accurate representation of the deflections, but to provide a
qualitative verification of their acceptability in the actual case. For the all other stress and deflection
analysis, a dynamic multiplication factor of 1.2 has been used. This is a relatively low dynamic
multiplication factor but is suitable for the purposes of this vehicle since the Shell Eco-Marathon
racetrack is a smooth professional racetrack and the vehicle will be travelling at a low maximum

DMT26| |p.27
DMT26 p.27
speed of 8 ms-1 (Simmons, 2019). The dynamic loading conditions would also involve centripetal loads
translated to the chassis, causing torsional stresses that have been modelled using FEA.

2.6.2.2 Simply Supported Beam Model of Chassis

As shown in Figure 19 below, the simply supported beam was modelled as being subject to a centrally
applied load of 40g since the location of the true centre of mass was approximately halfway between
the wheels as seen in Figure 18. Modelling the load as 40g assumes that the total load would be split
equally between either of the side sections. In reality, some load would be taken by the transverse

40g

1.55 m

Figure 19 - shear force and bending moment diagrams for simplified model

spaceframe members joining the two side sections, making this a conservative estimate. In this
loading case, the shear force diagram and bending moment diagrams are as shown.

The maximum bending moment is therefore at the location of the centrally applied load and equal to
PL/4. This value can be used to find the maximum stress experienced by the beam using Equation 6.

𝑀𝑦
σ= (6)
𝐼

Should the beam in this model be treated as a single carbon fibre tube of outer diameter 20 mm and
inner diameter 18 mm, the second moment of area would be 2.7 x 10-9 m4. Applying this to Equation
6 under the maximum bending moment condition, the stress is found to be 563.2 MPa, which
exceeds the tube yield stress of 153.6 MPa found from testing. Considering the spaceframe’s
structure (Figure 17) it was identified that, at the location halfway
between the wheels, the spaceframe’s side section is comprised of three a

members. The second moment of area was therefore recalculated using

the parallel axis theorem (Equation 7), with the beam modelled as three
carbon fibre tubes stacked vertically (Figure 20).
Figure 20 – structure cross-section

𝐼𝑥𝑥 = Σ( 𝐼𝑎𝑎 + 𝑎2 𝐴) (7)

DMT26| |p.28
DMT26 p.28
The adjusted second moment of area was 5.59 x 10-8 m4. When this was applied to Equation 6, the
maximum stress was found to be 27.2 MPa which is less than the yield stress of either the plain
carbon fibre tube (153.5 MPa) or the carbon fibre tube with a joint in it (71.8 MPa). This model could
therefore be used to find the deflection in the spaceframe side section using Equation 8. This relation
was derived by taking the deflection-moment relation and applying boundary conditions of zero
deflection and zero moment at the simple supports.

𝑊𝐿3
𝛿= (8)
48𝐸𝐼

The Young’s Modulus of the beam model was calculated to account for the fact that the spaceframe
side section contained several joints between the front and rear wheels. These joined tube sections
have a Youngs Modulus of 734.3 MPa whereas the plain tube section would have a Young’s Modulus
of 23.8 GPa. Since the stress in bending is along the length of the beam, the composite of plain and
joined sections can be modelled as shown in Figure 21.

σ σ

Figure 21 – material model for beam

For this case, the Young’s Modulus can be found as in Equation 9, where vjs is the volume ratio of the
joint sections and Ejs and Ept are the respective Young’s Moduli of the joint section and plain tube.

1
𝐸= (9)
𝑣𝑗𝑠 1 − 𝑣𝑗𝑠
+
𝐸𝑗𝑠 𝐸𝑝𝑡

Since there are 9 joints between the front and rear wheel locations at nodes 5/6 and 21/22
respectively as seen in Figure 17, the volume ratio of the joint sections can be found as shown, where
all joints are modelled as 30 mm long sections.

𝑃𝑙𝑎𝑛𝑒 𝑇𝑢𝑏𝑒 𝐿𝑒𝑛𝑔𝑡ℎ: 𝐽𝑜𝑖𝑛𝑡 𝑆𝑒𝑐𝑡𝑖𝑜𝑛 𝐿𝑒𝑛𝑔𝑡ℎ

(1550 − 9 × 30): (9 × 30)

1280: 270

This gives a vjs value of 0.174. Inserting this into Equation 9 gives a total Young’s Modulus of 3.68 GPa.
Applying this to Equation 8 gives a side section deflection of 0.15 m. Though this is a significant
deflection, it is less than the clearance gap between the chassis and shell of 0.5 m and since this
simply supported beam model is a gross simplification of the real situation, the true deflection would

DMT26| |p.29
DMT26 p.29
be acceptably small. A truer representation of the deflection experienced was estimated by applying
energy methods and treating the structure as a pin-jointed frame.

2.6.2.3 Pin-jointed Frame Model and Energy Method (AWP)

This method involved treating the spaceframe side section as a pin-jointed frame simply supported at
the wheel locations (Figure 22) and finding the tensions in each member before applying Castigliano’s
theorem to find the deflection at a node where a force is applied (Equation 10).

𝑁
𝜕𝑈 𝑇𝑖 𝐿𝑖 𝜕𝑇𝑖
𝛿𝑗 = =∑ (10)
𝜕𝑊𝑗 𝐸𝑖 𝐴𝑖 𝜕𝑊𝑗
𝑖=1
Z

Figure 22 - pin-jointed, simply supported model

To make it possible to find all the tensions in the spaceframe section, the transverse angles of its
members were ignored and it was instead treated as a flat truss structure, with member lengths equal
to the projection of the member lengths in the X-Z plane as found from the CAD model. The energy
method required that loads be applied at nodes. The total side section load of 40g was therefore
modelled as equivalent loads W1 and W2 at nodes 9/10 and 13/14 – the nodes adjacent to the centre
of mass location – as shown in Figure 18.

Tensions were first found for each member using the method of sectioning and applying equilibrium
to the whole structure as well as individual nodes. The contribution of each member to the
deflections at nodes 9/10 and 13/14 were then found in terms on W1 and W2 using Equation 10. Table
18 below summarises these results. The Young’s Modulus value used was simply that for a plain tube,
23.8 GPa, and the area of the tube cross section was calculated to be 5.97 x 10-5 m2.

Table 18 - deflection contributions of spaceframe members

Member Tension 𝝏𝑻 𝝏𝑻 𝑳 (mm) 𝑻𝑳 𝝏𝑻
(x10-8)
𝑻𝑳 𝝏𝑻
(x10-8)
𝑬𝑨 𝝏𝑾𝟏 𝑬𝑨 𝝏𝑾𝟐
𝝏𝑾𝟏 𝝏𝑾𝟐
1-5 / 2-6 0 0 0 556 0 0
3-7 / 4-8 0 0 0 308 0 0
5-7 / 6-8 -0.69W1-0.35W2 -0.69 -0.35 137 4.59W1+2.33W2 2.33W1+1.18W2
7-11 / 8-12 -2.52W1-1.28W2 -2.52 -1.28 475 212W1+108W2 108W1+54.8W2
7-9 / 8-10 2.61W1+1.33W2 2.61 1.33 492 236W1+120W2 120W1+61.3W2
5-9 / 6-10 0 0 0 475 0 0
9-11 / 10-12 W1 1 0 130 9.15W1 0
11-13 / 12-14 -2.52W1-1.28W2 -2.52 -1.28 522 233W1+119W2 119W1+60.2W2
9-13 / 10-14 1.28W1-1.45W2 1.28 -1.45 538 60.2W1-70.3W2 -70.3W1+79.6W2
9-15 / 10-16 1.28W1+2.69W2 1.28 2.69 713 82.2W1+173W2 173W1+363W2
13-15 / 14-16 -1.24W1-2.60W2 -1.24 -2.6 229 24.8W1+52.0W2 52.0W1+109W2
13-17 / 14-18 -0.45W1-0.95W2 -0.45 -0.95 171 2.44W1+5.15W2 5.15W1+10.9W2

DMT26| |p.30
DMT26 p.30
 17-19 / 18-20 -0.61W1-1.29W2 -0.61 -1.29 235 6.15W1+13.0W2 13.0W1+27.5W2
15-19 / 16-20 -0.19W1-0.38W2 -0.19 -0.38 177 0.45W1+0.90W2 0.90W1+1.80W2
19-21 / 20-22 -0.67W1-1.38W2 -0.67 -1.38 248 7.84W1+16.1W2 16.1W1+33.2W2
15-21 / 16-22 0.65W1+1.35W2 0.65 1.35 350 10.4W1+21.6W2 21.6W1+44.9W2

Summing the sixth and seventh columns gave the deflection at node 9/10 and node 13/14
respectively. The expressions for each of these deflections are shown in Table 19. It was calculated by
a moment balance that the 40g load at the centre of mass position could be represented as a W1 load
of 27.1g and a W2 load of 12.9g. Multiplying each of these values by 1.2, the dynamic load factor,
gave a W1 load of 32.5g and a W2 load of 15.5g Applying these to the expressions for each of the
deflections gave the deflection values shown in Table 19.

Table 19 - deflections found from Castigliano’s theorem

Node Deflection Expression (x10-6) Deflection (mm)
9/10 8.89W1 + 5.61W2 3.69
13/14 5.61W1 + 8.47W2 3.08

These deflection values vary from that found using the oversimplified simply supported beam model,
0.15 m, by two orders of magnitude. This proves the extent of the beam model simplification,
showing that it was certainly a suitable model to justify the acceptability of the actual deflections.

Applying the W1 and W2 values to the tension expressions found in Table 18, the greatest tension is
found to be 1034.4 N in member 7-9 / 8-10. This corresponds to a stress of 17.3 MPa, which is less
than the yield stress of either a plain or joined tube. This validates that no tube member will yield.

The longest spaceframe member is 9-15 / 10-16 which is 713 mm. Since the Euler buckling load is
inversely proportional to the square of the member length as shown in Equation 11, this member was
used to find a compressive load buckling limit of 1247.6 N, which offers a conservative limit for all
other tube members. The greatest compressive load in the spaceframe is found to be 864.6 N in
members 7-11 / 8-12 and 11-13 / 12-14. This validates that no tube member will buckle.

π2 𝐸𝐼
𝑃𝐸 = (11)
𝐿2

It should be noted that this pin jointed model is also a simplification. The lugs and significant layers of
wet layup wrapping at each of the joints may make them more restrictive than a pin joint model
would imply, with restrictive moments potentially being present. This would suggest that the
deflection values found are an overestimate. However, the Young’s Modulus values in the region of
the joints would be lower than the plain tube value. This has not been accounted for in the calculation
of the deflections and would suggest that these deflection values are therefore an underestimate.
Another significant simplification is the modelling of all the loads experienced by the spaceframe
section as point-loads at two nodes. The fact that some of the side section spaceframe’s members are

DMT26| |p.31
DMT26 p.31
angled in the transverse direction could also be considered when finding the tensions in each of the
members to find more accurate deflection values. The qualitative effect of this simplification is not
easily discernible. In a typical case, an FEA model could be used to find more accurate deflection
values and verify the hand calculations. However, this would require an exact model of the way in
which the material behaves. Since the carbon fibre material of the tubes and carbon wrapping results
in variable material properties and behaviours, such an FEA would offer deflection values that are no
more accurate than those found with this energy method. Nevertheless, member stresses found in
the ‘FEA Corroboration’ section highlight discrepancies due to the simplifications of this model.

2.6.2.4 Maximum Bending Stress in the Carbon Fibre Tubes

Even though the chassis could be modelled as a pin-jointed frame, some tubes will inevitably be
loaded in bending, mainly the tube connecting nodes 9 and 10 as well as the tube connecting nodes
13 and 14 together. These tubes are located directly under the driver’s body and so they will be
expected to be exposed to the greatest bending loads. From these two tubes, the most critical one
will be the tube connecting nodes 9 and 10 together as it will carry the load exerted by the driver’s
back (on the floor panel behind this tube) as well as the driver’s bottom (on the floor panel ahead of
this tube). These loads are expected to account for approximately two thirds of the driver’s weight, or
approximately 40 kg. However, it is important to note that the floor panels will be supported along all
four edges which means that this tube will only experience one quarter of this expected load,
resulting in approximately 10 kg of load on this tube. This was increased to 12 kg (or 117.72 N) for the
purposes of the stress analysis calculations, to account for the 1.2 dynamic multiplication factor.

This load will be distributed along the length of the tube as the panels will be attached to the tubes
using cable ties at regular intervals. The presence of internal lugs as well as extensive exterior wet lay-
up bonding at each end of the tube was more accurately modelled as built-in end conditions, resulting
in the bending moment diagram shown in
Figure 23. The bending moment is
expressed in terms of 𝑤 which is the
magnitude of the distributed load per
metre (in this case 261.6 Nm-1) and 𝐿
which is the length of the tube (in this
case 450 mm).
Figure 23 – bending moment diagram for pin-jointed, simply supported
model
Applying the central bending moment to
Equation 6, it was possible to determine that the maximum bending stress acting at the centre of the

DMT26| |p.32
DMT26 p.32
tube’s length would be 8.17 MPa. This is significantly lower than the failure stress of the tubes
obtained from the testing phase of 153.6 MPa.

2.6.2.5 Maximum Bending Stress at the Nodes

It was also important to determine the maximum stresses acting at the nodes to ensure that the wet
lay-up bonding of the tubes would be strong enough for this application. The tube connecting nodes 9
and 10 was also used for this analysis.

In this scenario, the largest bending moments would occur at the built-in ends, as shown in Figure 23
𝑤𝐿2
above, which could be expressed as 12
. This results in a bending moment of 4.41 Nm acting at each

node of this tube.

Each layer of cured carbon fibre lay-up was estimated to have a thickness of 0.3 mm. It was also
estimated from prototype testing that a minimum of four layers would be required per joint.
Therefore, the geometry of the joint was approximated as a cylinder, with internal diameter equal to
the external diameter of the tubes (20 mm) and external diameter of 22.4 mm given by the four
layers of wrapping. Using Equation 6, the bending stress at the joints was estimated as follows:

4.41 × 0.0112
𝜎= = 11 𝑀𝑃𝑎
𝜋(0.01124 − 0.014 )
( 4 )

This value obtained was significantly less than the failure stress of the carbon fibre joints which was
found to be 71.8 MPa through testing.

2.6.3 Finite Element Analysis Corroboration of Chassis Design (AWP)

Finite Element Analysis (FEA) tools could be used to obtain a general idea of the distribution of
stresses across the different chassis members during normal operating conditions.

However, there are major limitations to how informative these results would be:

• Results are only useful if they could be validated through the use of simpler hand calculations
or experiments.
• Strains and deflections require an exact model of the way in which a material behaves and
mechanical property descriptors such as Young’s modulus and Poisson’s ratio. These cannot
be accurately determined for the carbon fibre materials used in this project.

DMT26| |p.33
DMT26 p.33
The material choice has little effect on the stresses obtained through these software tools as these
values depend on the chassis geometry and loading conditions. Deflection results, however, would
not be accurate. For the purposes of this analysis, the tube material was set to aluminium 6063-T6 as
this was observed to result in accurate stress values
of simpler models compared to hand calculations.

A 3D model of the chassis was entered into the

software to create a mesh where loads and fixed
geometry could be applied, as shown in Figure 24.
Using the same joint numbering system as in Figure
17, nodes 5, 6 21 and 22 were selected as
immovable points to model them as pin supports Figure 24 – vertical loads applied to chassis during FEA
analysis
which accurately approximates the behaviour of
the wheel bearings at those locations.

The weight of the tubes was considered negligible for this analysis compared to the magnitudes of the
applied loads as summarised in the ‘Loading Condition’ section. A dynamic multiplication factor of 1.2
was applied to all loads. These were then distributed along the corresponding tubes as shown in Table
20. The result of the analysis is shown in Figure 25.

Table 20 - vertical loads applied

Load applied vertically Tube
downwards (N)
117.72 1-2, 9-10
58.86 5-9, 6-10, 5-6, 9-13, 10-14, 13-14
39.24 17-18, 13-17, 14-18
78.48 15-16, 14-16, 13-15

Figure 25 – vertical loading FEA analysis results

DMT26| |p.34
DMT26 p.34
An important result is the stress obtained at the centre of the tube joining nodes 9 and 10. As shown
in Figure 25, this tube exhibited a stress of 8.27 MPa at its centre which is very similar to the value
obtained using hand calculation shown previously of 8.17 MPa. Similarly, higher stress can be
observed in the contour plot at the ends of this tube, corresponding to the location of the carbon
fibre wet lay-up joints. This agrees with the assumption stated previously suggesting that these joints
may behave as built-in instead of pin-jointed.

Overall, the stresses exerted on the chassis are very low, on average having a value in the range of 10
to 20 MPa, as observed in Figure 25. These values are significantly lower than the failure stresses
obtained through testing, suggesting that this design could successfully withstand the loads expected
during the Shell Eco-marathon competition. Nonetheless, there are regions were higher stresses are
observed in the results of this analysis. The maximum stress was determined to be 61.0 MPa acting on
nodes 3 and 4. This value, while below the failure stress of the joints, was still considerable which led
to particular care being taken in the wrapping process of these particular joints, to extract the
maximum available strength from the carbon lay-up material.

A similar study was carried out to investigate the stresses generated by the torsional deflection of the
chassis due to the lateral cornering forces generated at the track. As mentioned in the ‘Steering
Design’ section of this report, the maximum lateral force that could be generated at each wheel was
324 N. This results in a maximum lateral force of 648 N for the two front wheels which, in turn,
corresponds to a total moment of 164.6 Nm acting on the plane of the front wheels (approximated by
the plane made up by the nodes 5, 6, 7 and 8). This load was applied to the mesh generated
previously and the results obtained are shown in Figure 26.

Figure 26 – torsional FEA analysis results

These results suggest that the maximum stress acting on the chassis would increase, as expected. The
maximum stress was now found to be 70.3 MPa on joint 4 which is also below the critical failure

DMT26| |p.35
DMT26 p.35
stress of the joints. Nonetheless, it is still considerably high. However, it is important to note that this
loading condition corresponds to the worst-case scenario in which the tyres are at their limit of
traction with the road. This is equivalent to the vehicle travelling at a speed of 8 ms-1 through a corner
of radius 8 m, which is very unrealistic according to discussions with the Powertrain DMT.
Nevertheless, every effort was made during the manufacturing plan to ensure that there would be
additional carbon fibre material applied at these critical locations during the wet layup process.

2.7 Manufacturing of Spaceframe

2.7.1 Chassis Assembly
In total the spaceframe consisted of 38 tubes components, total
length of 14.06 m, and 18 lugged inserts. The 15 easy composites ltd
1 m carbon fibre tubes were marked up and labelled using a ruler
and tape ready for cutting. The lugged inserts were hollow, solid,
plastic parts 3D printed in the Imperial College Advanced Hackspace.

Tubes were cut in the pit garage, using a hacksaw with the extractor Figure 27 – chassis alignment
using internal lugs
fan to collect dust produced immediately. Masks, goggles, lab coats and
gloves were worn to mitigate the health hazards as outlined in the risk assessments. Once the tubes
were cut and it was identified which tubes connected to which lugs, the surfaces of the lugs were filed
down to slot into tubes easily. The majority of the tubes slot on easily, however additional cuts were
made where a tube required an extreme angle to fit onto a lug. The assembled spaceframe is shown
in Figure 27.

2.7.2 Final Layup Process

The process outlined here is an optimised and sophisticated layup constructed after multiple
prototype joints were produced and tested. Firstly, the surfaces of all tubes within 5 cm from joints
were filed using sandpaper while wetted and then cleaned with acetone to increase the surface
adhesion and maximise the strength of the cured layup. The Easy Composites Ltd EL2 resin and
hardener was weighed and mixed in a ratio of 10:3 and steps 1 to 5 below were repeated for all joints
and left to cure for 24 hours.

1. Cut a strip of weave around 0.5 m long and paint fully with the
resin hardener mixture.
2. Begin wrapping around each tube member tightly and slowly,
moving in a longitudinal direction (Figure 28), covering at least 30
mm of each exposed tube and ensuring the fibres are kept
Figure 28 – wrapping motion of
straight and prevented from fraying. weave

DMT26| |p.36
DMT26 p.36
 3. Wrap a coated strip of narrow weave around the vertex
centre itself by using a figure of eight motion.
4. Repeat steps two and three until a minimum of 3 layers
are applied around the joint.
5. Cover in the composite shrink tape (Figure 29) and apply
heat gun gently at a distance of 20 cm for 5 seconds to
shrink the tape without breaking it.

Figure 29 – use of shrink tape during

manufacture

Figure 30 – new carbon fibre chassis (left) vs old steel chassis (right)

Figure 30 above shows the new carbon fibre spaceframe side by side with the previous entry chassis.
The new spaceframe weight was 2.25 kg, compared to the previous entry chassis weighing 9.65 kg.
With steering system and powertrain interface components, the floor panels and the bulkhead fitted,
the chassis was 5.6 kg, and therefore still met the target set in the PDS of being at least 10% lighter
than the previous entry.

DMT26| |p.37
DMT26 p.37
3. POWERTRAIN INTERFACES
3.1 Motor Mounting
The motor mount plate is shown in Figure 31. Where the motor foot mount bracket has longitudinal
slots that allow the motor position to be adjusted to tighten the chain before bolting, the mount plate
has slots in the transverse direction to allow for adjustment to correct chain misalignment.

Transverse slots for chain

A concern with the motor mount plate was its potential misalignment correction
to begin vibrating in resonance if a periodic force due to a Longitudinal slots for
chain tightening
mass out-of-balance in the motor excites the plate’s
natural frequency. This could potentially damage the rear
spaceframe or the chain. The natural frequency of the
plate was investigated by modelling the plate as
rectangular, with a length of 355 mm (its average length)
and a width of 168 mm. Since the plate was to be bolted
to the rear chassis members along its left and right edges Figure 31 – final motor mounting design

via collar pieces, with its top and bottom edges left free, the situation was modelled as in Figure 32,
giving the following boundary conditions, where Z is the displacement function.

1. Z = 0 along AC and BD
𝜕𝑍
2. 𝜕𝑥
= 0 along AC and BD
𝜕2 𝑍 𝜕2 𝑍
3. 𝑀𝑦 = −𝐷 [ +𝜈 ] = 0 along AB and CD
𝜕𝑦 2 𝜕𝑥 2
Free
A B
y

Fixed x Fixed

C D
Free
Figure 32 – motor plate model for vibration analysis

Applying these boundary conditions to the general form of the displacement function in a rectangular
plate, shown in Equation 12, gives the result for the natural frequency shown in Equation 13
(Steinberg, 2000). The plate stiffness 𝐷 is calculated as in Equation 14.

𝜕4𝑍 𝜕4𝑍 𝜕4𝑍 𝜕2𝑍

𝐷 [ 4 + 2 2 2 + 4 ] + 𝜌ℎ 2 = 0 (12)
𝜕𝑥 𝜕𝑥 𝜕𝑦 𝜕𝑦 𝜕𝑡

DMT26| |p.38
DMT26 p.38
 3.55 𝐷
𝑓𝑛 ≈ √ (13)
𝑏2 r

𝐸ℎ3
𝐷= (14)
12(1 − 𝜈 2 )
Applying the relevant material properties of aluminium with a plate thickness value, ℎ, of 4 mm and a
free edge length, 𝑏, of 410 mm gives a natural frequency of 8.14 Hz or 488.4 rpm. This is less than the
motor running speed of approximately 5730 rpm as confirmed with the Powertrain DMT team,
therefore proving the motor mount design to be safe.

3.2 Axle Mount Component

The rear axle was to be 135 mm in length, with 30 mm long 10 mm diameter threaded sections
extending from its ends. To mount the axle, a pair of tube-hook components as in Figure 33 were
conceptualised.
Spaceframe Member
Rear Axle
Shoulder

Nut
Threaded
End

Figure 33 – initial axle mounting concept

At one end, the component would slot into the spaceframe’s rear-most members, with its shoulder
sitting against the end, and methacrylate adhesive would be applied to keep it in place (VM100 Black
as recommended by Easy Composites). On the other end, a hook would sit on the threaded axle ends
and a nut would be tightened against the outer face to secure the axle.

Figure 34- vertical, angular, and longitudinal misalignment respectively

Potential misalignments of the spaceframe’s rearmost members due to inaccuracy in the tube joining
and assembly method are shown in Figure 34. The conceptualised axle mount components would be
able to overcome vertical or longitudinal misalignment by the relative adjustments shown in Figure
35.

DMT26| |p.39
DMT26 p.39
 Figure 35- Adjustment methods to correct for longitudinal and vertical misalignment respectively

To overcome longitudinal misalignment, the axle mount component on the side for which the rear
spaceframe member is relatively short would be extended out of the tube by an appropriate amount
before methacrylate adhesive is applied (such that the tube end would no longer sit against the
shoulder). To overcome vertical misalignment, the hook part of the components would be made long
enough such that the nut could still be tightened against its outer face without the hook having to sit
flush with the threaded end.

Nevertheless, the conceptualised axle mount component could not be adjusted to overcome angular
misalignment that might occur. Table 21 therefore summarises two considered adjustments to the
axle mounting method to mitigate against angular misalignment and their relative merits.

Table 21 - mounting axle component adjustment options

Proposed Addition of short, straightening spaceframe members. Component made in two parts – a tube
Adjustment fitting part that bolts to a hook piece.

Explanation Once the spaceframe is assembled, short members Each of the hook pieces can be slotted into
would be joined to the rear-most spaceframe members, the tube fitting parts and bolted at
ensuring that they are straight and parallel. appropriate angles such that the hook
pieces are straight and parallel.

Sketches Hook piece

Tube fitting
part

Pros - Vertical and longitudinal misalignment mitigation - Vertical and longitudinal misalignment
methods proposed in Figure 35 can still be employed. mitigation methods proposed in Figure 35
can still be employed.
- Also offers extra mitigation against longitudinal
misalignment through length of straightening members. - Easily adjustable to mitigate angular
This would be useful in case this misalignment is so misalignment completely.
significant that extension of the component out of the
tube as in Figure 35 is insufficient.

Cons - Since the angle of the joints that need to be made - It would be preferred for the axle mount
would be dependent on the observed angular components to be manufactured using CNC
misalignment and not on the intended design, creating to maintain accuracy and symmetry
the lugs required for these joints would prove between each of the pair. Splitting the
challenging. component into two will mean that the axle
mounts will account for four CNC requests

DMT26| |p.40
DMT26 p.40
 - The joining of these short members may lead to rather than two. Since several steering
misalignments just the same as any other joint in the system components also had to be made by
spaceframe, which was the cause for misalignment CNC milling, this would likely mean that
concerns in the first place. However, since the some of these axle mount CNC requests are
misalignment is now dependant on only this joint, the rejected, meaning parts would have to be
likelihood is lower as misalignment wouldn’t aggregate manufactured by hand. This is potentially a
across all the spaceframe joints. source of inaccuracy and asymmetry, which
itself has the potential to result in
- The region near the axle mount would experience the misalignment.
reaction force of the rear wheel as well as shock forces.
Therefore, the shortness of the tube member and the
weakening effect of the joint may result in yielding
stresses here.

- The rear-most point in the spaceframe was designed to

be a node to improve the stress distribution. Adding
these short members would therefore negate this
consideration, worsening the stress distribution as the
forces experienced at the axle mounts are cantilever
loads.

It was reasoned that the adjustment of the axle mount concept to be made in two parts would be the
best option to mitigate against angular asymmetry in the rear spaceframe members. This design was
further developed, and the final design was as shown in Figure 36. The length of the tube fitting
before its shoulder was extended to 60 mm to ensure a significant surface could be painted with
methacrylate adhesive and a strong bond could be present with the inside of the tube. This also
increased the component’s ability to mitigate against longitudinal misalignment.

Figure 36 – final design of axle mounting component

Due to a lack of CNC machine time availability, the tube fitting parts were hand manufactured from 25
mm diameter, 85 mm long cylindrical mild steel bars. The challenge of creating the rounded face was
overcome using a radial cutter with the part oriented vertically in a milling machine. Though a
completely smooth curve profile on the tube fitting face could not be created through this process as
seen in Figure 37, the curve was sufficient for the hook piece to experience no impedance of angular
adjustment due to sharp corners.

Rounded face that

presented manufacturing Figure 37 – manufactured axle
mounting component
challenge

DMT26| |p.41
DMT26 p.41
4. STEERING DESIGN
The steering system of this project can be divided into three areas of investigation that would inform
the design process:

• Steering geometry and variable angles.

• Interface between steering input and front wheels
• Driver control and actuation method

4.1 Steering Geometry and Variable Angles

4.1.1 Introduction to Adjustable Steering Angles
The performance and handling of a steering system depends on several design parameters:

• Kingpin Inclination Angle – refers to the angle between the vertical axis of the wheel when
viewed from the front and the pivotal axis of the wheel (Figure 38). The distance between these two
axes on the surface of the track is the ‘Scrub Radius’.

• Camber Angle – refers to the angle between the vertical axis of the wheel when viewed from
the front and the perpendicular axis to the ground (Figure 39).

• Toe angle – refers to the angle between the direction in which the wheel is pointing when
viewed from above and the centreline of the car (Figure 40).

• Caster angle – refers to the angle between the pivotal axis from the wheel when viewed from
the side and the perpendicular axis passing through the tyre’s contact patch to the ground (Figure
41). The distance between these two axes on the surface of the track is the ‘Mechanical Trail’.

Figure 38- Schematic showing kingpin Figure 39- Example schematic Figure 40- Schematic Figure 41- Schematic showing
inclination angle (referred to as showing 4° of negative camber showing toe angle (Seward, caster angle (Seward, 2014).
‘steering axis inclination’) and scrub (Seward, 2014). 2014).
radius (Seward, 2014).

Different configurations are preferable depending on the operating conditions of the vehicle.
Therefore, the objective of fuel-efficient operation informed target angles.

Many of the handling characteristics of the vehicle will depend on the type of tyre used. There are a
range of different tyre properties such as the rubber compound used or the internal structure of the

DMT26| |p.42
DMT26 p.42
tyre that can impact the handling characteristics of the vehicle by affecting contact patches, slip
angles, camber thrust, among other tyre-track interface behaviour (Santing et al., 2007). Studying
these effects falls outside the scope of this project. However, it was decided that the steering
mechanisms should provide sufficient adjustability to allow for optimum toe and camber angles to be
selected, through testing, for any chosen set of tyres.

Other design decisions such as caster angle and kingpin inclination are not normally adjustable so
they were thoroughly investigated before deciding on the most appropriate choice.

4.1.2 Kingpin Inclination

Looking at the work carried out by Milliken & Milliken (1995), it was discovered that the greater the
kingpin inclination angle, the more the vehicle would be raised when the steering wheel is turned
because the pivoting axis is not perpendicular to the ground. This increases the effort needed by the
driver to turn the vehicle, as they are effectively exerting a force against a component of the weight
of the car, but it also helps to centre the wheels during low speed turning. However, increasing the
kingpin inclination can add positive camber to the vehicle when turning which is undesirable, making
the steering feel unresponsive and causing unnecessary loads to compress the sidewall of the tyre.

The kingpin inclination also has the effect of providing a design method for altering the scrub radius of
the tyre. Reducing this scrub radius reduces the tyre wear by minimising friction between the tyre and
the track and it also reduces the moment arm of any braking forces acting on the wheel, resulting in
more stable steering under braking.

Minimising the scrub radius (ideally, making it as close to zero as possible) means that the kingpin
inclination that can be achieved will largely depend on the clearance available between the wheel hub
and the kingpin which is often dependent on other components such as the brake disc, brake caliper
or tie-rod mounting.

4.1.3 Caster
The caster angle of the front wheels is the main factor which can aid the self-centering of the wheels
after turning. The distance between the contact patch of the tyre and the rotating axis (kingpin axis)
of the wheel is referred to as the mechanical trail and it effectively acts as the moment arm for all the
lateral forces that are transmitted between the road and the car during cornering.

As the car is effectively undergoing circular motion as it drives around a corner, the lateral forces due
to friction are transmitted to the wheel, generating a moment about the kingpin axis. Therefore,
increasing the mechanical trail has the effect of increasing the moment generated at the wheels
causing the front wheels to have a greater propensity to return to the straight ahead position. This, as

DMT26| |p.43
DMT26 p.43
stated by Milliken & Milliken (1995), can be beneficial as it results in more stable and smooth steering
at the corner exit but it also results in heavier steering entering the corner as the driver has to
effectively counteract the increased moment to turn the front wheels.

4.1.4 Steering Angles Selection

Initial research carried out by the team revealed that the ideal camber and toe angles for a fuel
efficiency race should be 0° (Santing et al., 2007). This is in contrast to other set-up configurations
used in most racing events were maximum corner speed is usually the determining factor. In this
case, the main priority was to reduce all sources of friction and this was found to be achieved by
minimising the toe and camber angles.

The kingpin inclination angle should ideally be selected such that a scrub radius of zero is achieved.
Based on initial estimations of clearance for other components around the wheel hubs, the maximum
kingpin inclination angle would be approximately 17°, measured from the vertical.

Since the kingpin inclination angle has the effect of increasing the camber angle of the wheel while
turning, as mentioned in the ‘Kingpin Inclination’ section, it is important to select a relatively small
kingpin inclination angle so that the behaviour of the car is relatively constant and predictable.
However, it should not be so small that the advantages of implementing a kingpin inclination angle
become insignificant.

The expression for the change in camber angle while turning, due to the kingpin inclination, is given
by Seward (2014) as follows:

Δ𝛾𝑘 = 𝜃𝑘 + cos−1(sin 𝜃𝑘 cos 𝛿) − 90 (15)

In this expression, Δ𝛾𝑘 is the change in camber due to kingpin inclination, in degrees, 𝜃𝑘 is the chosen
kingpin inclination angle, in degrees, and 𝛿 is the steering angle (the angle of the front wheel relative
to the straight ahead position), in degrees.

Seward (2014) also suggests that the caster angle has an effect on the camber angle of the wheels
while turning. However, while the kingpin inclination angle has the effect of adding undesirable
positive camber angle, the caster angle has the effect of adding negative camber angle. Therefore,
the caster angle of the wheel can be selected so that it counteracts the adverse effects of the kingpin
inclination angle while turning.

The expression for the change in camber angle while turning, due to the caster angle, is given by
Seward (2014) as follows:

Δ𝛾𝑐 = cos −1 (sin 𝜃𝑐 sin 𝛿) − 90 (16)

DMT26| |p.44
DMT26 p.44
In this expression, Δ𝛾𝑐 is the change in camber due to caster angle, in degrees, 𝜃𝑐 is the chosen caster
angle, in degrees, and 𝛿 is still the steering angle, in degrees.

Therefore, a total change in camber while turning, Δ𝛾𝑡𝑜𝑡𝑎𝑙 , in degrees, can be calculated by adding
the two expressions above.

Δ𝛾𝑡𝑜𝑡𝑎𝑙 = Δ𝛾𝑘 + Δ𝛾𝑐 (17)

This car was designed for an extreme efficiency race. This meant that, ideally, the camber angle
should always be zero as this results in the minimum amount of rolling resistance. Therefore, a
combination of kingpin inclination angle and caster angle needed to be selected so that the resulting
change in camber angle while turning was as close to zero as possible.

Practical design considerations that became apparent further along in the ideation stage of the
steering system resulted in the selection of rod-ends to be used as the upper and lower pivot points
between the chassis and the wheel hubs. Rod-ends were selected instead of spherical bearings
mounted in bearing housings due to their lighter weight, smaller dimensions and ease of mounting
without additional fasteners. The line joining these upper and lower pivot points is the kingpin axis.
However, these rod-ends, which were to be obtained from RS Components, were commonly found to
have a maximum allowable misalignment of ±13°. Taking this into consideration, the kingpin
inclination angle was selected to be 12° which is the highest value that would still allow 1° of
misalignment due to expected manufacturing inaccuracies.

A comprehensive analysis was then carried out to observe how different combinations of caster angle
and the chosen kingpin inclination angle of 12° would affect the steering system for different steering
inputs, the results of which are summarised in Table 22. It was found that a caster angle of 4° resulted
in a more stable value of Δ𝛾𝑡𝑜𝑡𝑎𝑙 which remained close to zero and negative across all steering angles
(positive camber is undesirable as previously mentioned).

Table 22 - caster angle analysis

Steering Input/ 𝚫𝜸𝒕𝒐𝒕𝒂𝒍 / ° for 𝚫𝜸𝒕𝒐𝒕𝒂𝒍 / ° for 𝚫𝜸𝒕𝒐𝒕𝒂𝒍 / ° for 𝚫𝜸𝒕𝒐𝒕𝒂𝒍 / ° for 𝚫𝜸𝒕𝒐𝒕𝒂𝒍 / ° for
° 2° caster 3° caster 4° caster 5° caster 6° caster
0 0.00 0.00 0.00 0.00 0.00
5 -0.13 -0.22 -0.30 -0.39 -0.48
10 -0.16 -0.34 -0.51 -0.69 -0.86
15 -0.11 -0.37 -0.63 -0.89 -1.14
20 0.04 -0.30 -0.64 -0.99 -1.33
25 0.28 -0.14 -0.57 -0.99 -1.41
30 0.61 0.11 -0.39 -0.89 -1.39

A summary of the selected steering angles is shown in Table 23.

DMT26| |p.45
DMT26 p.45
Table 23 - Selected steering angles
Angle Value/ °
Toe 0
Camber 0
Caster 4
Kingpin Inclination 12

It was expected that other factors such as flexing of the chassis under load as well as manufacturing
tolerances would result in deviations from the desired toe and camber values. This meant that
methods for adjusting these angles were worth implementing in this project.

4.1.5 Steering Linkages Geometry

Different steering geometries were investigated for this project. This refers to the geometric
arrangement of the steering linkages which connect the two front wheels together which can have an
effect on the behaviour of the car around corners. Ackermann and Reverse Ackermann geometries
were investigated.

• Ackermann: When the vehicle is driving around a corner, each of the front wheels travel
around an arc of different radius, due to the track width of the vehicle. The Ackermann
geometry allows for the inner wheel to rotate more than the outside wheel so that both
wheels are pointing in the correct direction, minimising friction between the wheel and the
track surface.
• Reverse Ackermann: This is the opposite of Ackermann geometry in the sense that it forces
the outside front wheel to rotate by a greater angle than is necessary. This is a set-up often
used in high-performance vehicles and is a compromise between improved turning at high
speed and added friction between the tire and the track surface. It is clear that this set-up is
more beneficial for high-performance competitions which require high corner velocities.

It was decided that the Ackermann geometry would be the best choice for this type of vehicle
considering that the main objective was to maximise efficiency. Furthermore, the implementation of
Ackermann geometry only depends on the location of the pin joint of the tie-rod connecting the front
wheels to each other which means that it does not add any unnecessary complexity or additional
components to the design, which is another reason why it was favoured.

4.2 Interface Between Steering Input and Front Wheels

Three methods were investigated for transferring the steering input from the driver to the front
wheels. These were based on methods commonly used in the automotive industry as well as the
previous experience of the Shell Eco-marathon team. An overview and description of each mechanism
is provided in Table 24.

DMT26| |p.46
DMT26 p.46
Table 24 - considered steering input to front wheel mechanisms
Proposed Idea Description Advantages Disadvantages
1. Rack and Pinion Consists of a Pinion Uses components already Sensitive to misalignment
attached to the steering available from a previous and can be unreliable.
column and a rack DMT project. Provides Heavier than other options
connected to the front greater choice for steering due to the increased
wheels. ratio as only the pinion gear number of metal
would need to be replaced. components.
2. Tie-rods on Steering Consists of connecting the Implemented in the Wheel angle may not
Column tie-rods from the wheel previous generation Shell change linearly with
hubs directly to the Eco-Marathon car and is steering input angle –
steering column (or lighter than other options depends on the geometry
through the use of a due to the smaller number of the Pitman arm.
Pitman arm – a lever of metal components. It is
which extends from the also more reliable as it is a
steering column which is pin-jointed mechanism,
rigidly connected to the which is less sensitive to
column). misalignment.

3. Single lever Consists of a lever Visibility may be improved Slightly more complex than
mechanism mounted at due to steering system Option 2 due to complex
the side of the car and mounted at the side of the linkage geometry between
transferring the steering chassis. steering column and front
input from the driver to wheel. Little space at
just one of the front either side of the driver so
wheels directly. A tie-rod this may interfere with
is then used to connect other components.
the two front wheels
together.

Using the PDS as a guideline, each idea was evaluated using a decision matrix with the following
categories: cost, weight, reliability, ease of mounting to chassis and ease of manufacture.

Weightings were assigned to each category from 1 to 5, with 5 signifying the most important category
for this application. Each idea was then ranked from 1 to 3, with 3 being assigned to the idea that best
fulfils the requirements of each category. The decision matrix is shown in Table 25.

Table 25 - decision matrix for steering input to front wheel mechanism

Option 1 Option 2 Option 3
Weighted Weighted Weighted
Criteria Weighting Score Score Score Score Score Score
Cost 3 1 3 2 6 3 9
Weight 2 1 2 3 6 2 4
Reliability 5 1 5 3 15 2 10
Ease of
Mounting to
Chassis 1 3 3 2 2 1 1
Ease of
Manufacture 4 3 12 2 8 1 4
Total 1: 25 Total 2: 37 Total 3: 28

DMT26| |p.47
DMT26 p.47
Using the results from the decision matrix shown in Table 25, it was decided that Option 2, consisting
of mounting the tie-rods to the steering column, would be the best choice for this application.

4.3 Driver Control and Actuation Method

A very similar process was carried out for the driver control and actuation system for the car, which
refers to the method through which the driver would interact with the steering system of the vehicle.
Four ideas were proposed and they were compared against the requirements described in the
Product Design Specification to then select the most appropriate choice through the use of a decision
matrix. The options explored are described below.

Table 26 - actuation methods

Proposed Idea Description Advantages Disadvantages
1. Steering Wheel Most familiar method More ergonomic than Needs to be purchased
for controlling the other options and or careful
steered wheels of a allows the driver to consideration given to
vehicle. comfortably apply the its manufacture due to
required steering its complexity. Poor
torque. visibility straight ahead
due to the low driving
position and it may be
heavier than other
options.
2. Handlebars Implemented in the Relatively cheap and May provide
previous vehicle and is simple to manufacture. uncomfortable
similar to the It is also the best ergonomics, especially
handlebars in a option for easily during turning.
bicycle. attaching other
components such as a
throttle or brake
levers.

3. Lever Handle Consists of a single Lightweight due to the May be difficult to

lever stick that would reduced number of operate and tilt left
be used to control the parts. and right in the
wheels. It would be confined space of the
better suited for the driver’s cockpit.
‘Single Lever’ steering
mechanism but it can
be adjusted to the
requirements of the
‘Tie-rods on Steering
Column’ mechanism
chosen.
4. Angled Handlebars Similar idea as Option Ergonomics would be Visibility would be
2 but different improved by angling slightly reduced
configuration of the handlebar and compared to the
handles. offsetting the pivot ‘Handlebar’ option.
point vertically so that
the driver has more
space to operate the
steering.

DMT26| |p.48
DMT26 p.48
Another decision matrix was implemented at this stage, although the five categories chosen to
compare each option were now changed to the following: cost, weight, ergonomics, ease of
manufacture and visibility

Weightings were also assigned to each category, from 1 to 5, and each idea given a rank from 1 to 4.
The results are shown in Table 27.

Table 27 - decision matrix for actuation methods

Option 1 Option 2 Option 3 Option 4
Weighted Weighted Weighted Weighted
Criteria Weight Score Score Score Score Score Score Score Score
Cost 1 1 1 3 3 4 4 2 2
Weight 2 1 2 3 6 4 8 2 4
Ergonomics 5 3 15 1 5 2 10 4 20
Ease of
Manufacture 3 1 3 3 9 4 12 2 6
Visibility 4 2 8 4 16 1 4 3 12
Total Total Total Total
1: 29 2: 39 3: 38 4: 44

Option 4, consisting of the angled handlebars, would be the best choice for this application. One of
the main reasons for this is due to the improved ergonomics which is of great importance given the
very small space available for the driver inside the car, especially in endurance competitions.

In order to reduce the weight of the car, it was decided to manufacture the steering interface from
carbon fibre. However, an angled design would have been difficult to implement without proprietary
moulds, increasing the complexity and cost of this option. Similarly, the penalty on visibility was made
clear once the complete vehicle was developed using computer aided design.

Ultimately, the decision was made to implement the next best option from the decision matrix
analysis. This was Option 2 which consisted of a straight handlebar design. The team were able to
reuse the previous prototype Shell Eco-marathon entry’s carbon fibre handlebar steering column with
minimal modification required or additional cost.

4.4 Steering Design Process

4.4.1 Bearing Selection and Design Considerations
At this stage in the design process, it was already agreed with the supervisor and the team principal
that the wheels of the previous generation Shell Eco-Marathon vehicle would be reused for this
project. Therefore, the bearings already mounted on their wheel hubs would be reused and the only
other bearings that would need to be selected would be those that allow the wheel spindle assembly
(consisting of all the components that are mounted directly onto the wheel) to rotate about the

DMT26| |p.49
DMT26 p.49
kingpin axis. The bearing selection process shown below corresponds to the analysis carried out for
these kingpin axis bearings.

4.4.1.1 Design Criteria

It was important to determine an estimate for the angular velocity of rotating components in the
steering system, even if these would be subject to partial rotations as opposed to complete
revolutions. A number of conservative estimates were made based on information previously
collected about the track, the regulations and previous experiences.

It was determined that a good estimate for the target turning radius for this vehicle would be 6
metres. Taking into account the selected wheelbase of 1.55 m and a track width of 500 mm, a
steering angle, 𝜃, of approximately 15° on the outside wheel was calculated from Equation 18.

𝑤ℎ𝑒𝑒𝑙𝑏𝑎𝑠𝑒
𝜃 = sin−1 ( ) (18)
𝑡𝑢𝑟𝑛𝑖𝑛𝑔 𝑟𝑎𝑑𝑖𝑢𝑠 𝑜𝑓 𝑤ℎ𝑒𝑒𝑙

This corresponds to a maximum steering angle of 16.3° on the inside wheel as its turning radius is
smaller due to the Ackermann steering geometry selected. However, to consider any potential excess
steering made by the driver, a conservative maximum steering angle of 20° at the wheels was used
for calculations.

The driver’s reaction time was estimated to allow the driver to cover this turning angle in a period of
0.5 seconds. Therefore, the equivalent rotational speed of the wheels rotating about their kingpin axis
would be 6.67 revolutions per minute.

This angular velocity would provide a design criterion for the bearings that would support the kingpin
axis. However, this value was extremely low as a requirement for rolling element bearings - any rolling
element bearing would easily satisfy this requirement. Therefore, other considerations such as load
cases as well as efficiency concerns would be more appropriate selection criteria for these bearings.

4.4.1.2 Load Case

The overall static axial load on the entire vehicle was estimated to be 110 kg, corresponding to 50 kg
for the vehicle (including the chassis, steering, powertrain, pedal box and shell) as well as 60 kg for
the driver. Note that at the time of calculation, the chassis design was still in its early stages, leading
to a conservative worst-case estimate for its mass. A safety factor of 1.2 was applied to this value, as
required in the Product Design Specification, resulting in a total static weight of 132 kg to be used as a
design consideration.

This value was initially divided up equally across the three wheels to provide an initial estimate of 432
N of axial load acting on each wheel assembly. However, in each of the front wheel assemblies, the

DMT26| |p.50
DMT26 p.50
wheel spindles were expected to be connected to the chassis by two bearings, one at the top and
another one at the bottom of the kingpin axis. Therefore, the total static axial load on each bearing
would be 216 N.

The dynamic axial load on each bearing was also estimated to be 216 N. This value was chosen to be
the same as for the static load case because the team learned that the competition would be held at a
professional test track consisting of a relatively smooth surface as well as no significant elevation
changes and as such a dynamic multiplication factor of 1.2 was considered appropriate.

The dynamic radial load on each bearing was estimated using information provided to this team by
the Powertrain DMT team. They estimated that the acceleration produced by the motor would be
approximately 0.3 ms-2. Using Newton’s 2nd Law of Motion, the force required to produce this
acceleration was 39.6 N that would correspond to 13.2 N at each wheel assembly and, therefore, only
6.6 N at each bearing.

Similarly, the maximum tractive deceleration force that could possibly be achieved would correspond
to the maximum friction that could be generated between the tyres and the track surface. Using an
estimated value of 0.75 for the friction coefficient and a total static vehicle weight of 132 kg, this
corresponds to a deceleration force of 971.2 N that would be distributed across all three wheels
resulting in 324 N at each wheel and, therefore, 162 N at each bearing. Comparing the maximum
acceleration and deceleration forces at each bearing, it was possible to say that the maximum radial
load to be expected at each bearing would be 162 N.

The maximum lateral force was also calculated and, similarly to the deceleration force, it would occur
when the vehicle is at the point of sliding as it travels around a turn. The analysis results in the same
value of 971.2 N of lateral force on the vehicle, corresponding again to 162 N of force at each bearing.

A summary of the maximum value of each type of load acting at each bearing is shown in Table 28.

Table 28 - loading on bearings

Load Type per Bearing Maximum Value/ N
Static Axial Load 216
Dynamic Axial Load 216
Radial Load 162

4.4.1.3 Bearing Selection

The bearing type selection chart from ‘Mechanical Design Engineering Handbook’ (Childs, 2019)
shown in Figure 42 was used to select bearings that would satisfy the requirements stated previously.

DMT26| |p.51
DMT26 p.51
This low speed and low load application lies within the region for rubbing plain bearings. The
operating conditions also lie within the 5 mm diameter region of the graph, in Figure 42. Therefore, as
a starting point, bearings with 5 mm internal diameter were investigated.

The final decision was made to use rubbing plain bearings, agreeing with the analysis and selection
method using the chart in Figure 42, in the form of spherical bearings mounted on rod-ends. These
would allow for misalignment between the top and the bottom kingpin axis mounting points, as
mentioned in the ‘Steering Angles Selection’ section of this report. The best option found from
available suppliers were the LDK POS 5EC series of rod-ends as these were found to have the desired
dimension of 5 mm internal diameter as well as the correct misalignment of 12° required for the
kingpin axis inclination implemented in this design.

Figure 42 – maximum load against speed chart for spherical bearings

4.4.1.4 Internal Diameter Bearing Check

It was important to check the required strength for the material that will act as a shaft through both
bearings to make sure that a maximum diameter of 5 mm would be sufficient.

Design decisions that were made further along in the embodiment phase of the design stage resulted
in the implementation of a threaded rod that would act as the kingpin axis and would, therefore, pass
through the top and bottom rod-ends chosen previously. This rod would be mostly loaded in shear
due to the lateral forces acting on the car as well as the acceleration and deceleration forces
previously mentioned. Therefore, research was carried out into the shear failure mode of bolts,
representing the threaded rod selected, and it was found that their shear strength was approximately
60% of the ultimate tensile strength of the material which the bolt is made from (Fastenal, 2009).

DMT26| |p.52
DMT26 p.52
The maximum force acting on the bolts in shear would be 162 N. Assuming that the strongest
available bolt can be sourced, this corresponds to a Class 12.9 bolt with an ultimate tensile strength of
1220 MPa and therefore a shear strength, 𝜏, of 732 MPa. Using Equation 19, it was determined that
the diameter, 𝐷, required for the bolt would be 0.5 mm.

𝑆ℎ𝑒𝑎𝑟 𝐹𝑜𝑟𝑐𝑒 𝑆ℎ𝑒𝑎𝑟 𝐹𝑜𝑟𝑐𝑒

𝜏= = (19)
𝐴𝑟𝑒𝑎 1 2
4 𝜋𝐷

If a lower grade bolt is used, which is much more readily available, it would correspond to Class 8.8
properties with an ultimate tensile strength of 800 MPa and a shear strength of 460 MPa. The
diameter required for the load case being studied would be 0.66 mm.

In conclusion, even if the lower grade bolts were used, any practical diameter would suffice for this
application. This analysis, which highlighted the sufficient strength of 5 mm diameter bolts, made it
possible to select M5 or bigger bolts for any mounting requirements throughout the vehicle, with the
selection being more dependent on practicality and available components instead of strength.

4.5 Provisional Steering Assembly Design

4.5.1 Wheel Spindle and ‘L’ Bracket Assemblies
Instead of using a decision matrix approach, the iterative nature of this design project meant that it
was more beneficial to begin with a basic idea that would then be improved until it satisfied all the
design specifications.

The provisional design chosen at this stage of the development process is shown in Figure 43.

Figure 43a – wheel spindle assembly sketch Figure 43b – ‘L’ bracket assembly sketch

This design consists of two separate subassemblies at each front wheel which are connected by the
rod-ends. The subassemblies are as follows.

DMT26| |p.53
DMT26 p.53
 • Wheel Spindle Assembly: This subassembly will be mounted to the wheels of the car and,
therefore, consists of all the components which are required to turn left and right with the
wheel. This consists of a main plate that will provide the general structure onto which other
components will be attached, two angled brackets which provide the correct kingpin axis
inclination, an Ackermann arm that allows the appropriate Ackermann steering geometry to
be achieved, a wheel shaft which is stationary and is machined down to the appropriate
diameter of the bearings already mounted on the wheels, and appropriate mounting for the
brakes which were selected by the Powertrain DMT. All components would be made of
aluminium to reduce weight.
• ‘L’ Bracket Assembly: This subassembly will be mounted semi-rigidly to the chassis and will
provide mounting points for the wheel spindle subassembly to connect via the rod-ends. It
consists of two chassis mounting plates that will have mounting points to connect to the
chassis tubes as well as clearance slots for camber bolts to allow the camber to be adjusted,
and an ‘L’ shaped bracket that will have appropriate mounting points for the rod-ends that
will be horizontally offset from each other to provide the correct kingpin axis inclination and
caster angles. All components would be made of aluminium to reduce weight.

It was decided that for the wheel spindle assembly, all components should be made separately and
then mechanically fastened to the main plate through the use of M5 bolts and nuts. This method
presents a weight penalty compared to a unibody design, but it is cheaper to manufacture using in-
house methods and also provides greater flexibility for further development of each component. CNC
was ultimately implemented to achieve the desired angle for the angled brackets using the facilities
available at the Student Teaching Workshop.

4.5.1.1 Design Considerations

An important design choice was the thickness of all the components to ensure that they could
withstand all the forces expected. An example of a critically loaded component was the angled
bracket that would effectively act as a built-in cantilever beam. The driver and chassis loads would be
transferred by the ‘L’ bracket rod-ends and be exerted on the angled brackets by spacers mounted on
the kingpin axis bolt.

Initially, a thickness of 3 mm was chosen for the

cantilever section of the bracket to minimise the
total weight. An approximate width of 20 mm was
chosen to provide sufficient support for the kingpin

Figure 44 – angled bracket dimensions

DMT26| |p.54
DMT26 p.54
axis spacers and a total cantilever length of 20 mm was used as an initial estimate for the length of
the angled brackets, as shown in Figure 44.

As mentioned previously, a maximum axial load of 216 N was expected to act vertically at each wheel.
This provided a force component of 211 N perpendicular to the surface of the cantilever angled
bracket. The bending moment calculations and bending stress analysis were carried out as follows
using Equation 6, where the bending stress, 𝜎𝑏𝑒𝑛𝑑𝑖𝑛𝑔 , is a maximum.

𝑀 = 211(0.02) = 4.23 𝑁𝑚

𝑀𝑦 4.23(0.0015)
𝜎= = = 141 𝑀𝑃𝑎
𝐼 0.02 × 0.0033
( )
12

For a 3 mm thick piece of aluminium of the given dimensions, the maximum bending stress was found
to be 141 MPa. This was too high for the grade of aluminium available from the ME Stores (6082-T6
grade) which had a yield strength range from approximately 180 MPa to 230 MPa. The same analysis
shown above was repeated for a thickness of 4 mm and the maximum bending stress was found to
decrease to a value of 79.3 MPa which was acceptable. It was therefore decided that a minimum
thickness of 4 mm would be used for all components of the steering system. A 2.5 mm fillet radius
was added at the base of all cantilevered components to mitigate the effects of stress concentration.

4.5.2 Brake Caliper Mounting Arm

Another critical feature that needed to be included in the design of the steering system was the
mounting for the front brake calipers. Therefore, it was necessary to design the main plate of the
wheel spindle assembly to include appropriate spacing for the mounting of an additional bracket, if
necessary, to securely hold the brake calipers. The specific selection of the brake caliper parts would
be the responsibility of the Powertrain DMT.

The coordinates of the mounting points of the brake calipers on the main plate relative to the axis of
the wheel depends on the diameter of brake disc used. The brake disc is the component that is
attached to the wheel where the caliper exerts the braking force. Bicycle wheels were to be used in
this car which meant that it was necessary to investigate which type and size of bicycle brakes were
available from convenient suppliers. This was particularly important considering the modification to
the rules and regulations that was introduced for the Shell Eco-Marathon competition in 2020 which
required hydraulic brakes to be used throughout the vehicle (Simmons, 2019). It was found that brake
rotors were supplied in standard dimensions of 160 mm, 180 mm and 203 mm in diameter.

This was compared to the previous generation car which was found to successfully implement ‘TRP
14’ 160 mm brake rotors. Therefore, after discussion with the team principal, it was decided that this

DMT26| |p.55
DMT26 p.55
vehicle would also use the same dimension for the rotors in an effort to reduce the total weight of the
vehicle. This decision was also discussed with the Powertrain DMT who confirmed that these
dimensions would provide sufficient braking capacity for the dynamic conditions that this car would
experience. Brand new ‘Avid HS1’ 160 mm brake rotors were found in the Pit Garage, that had
significantly less wear than the old parts, were perfectly flat without any warping and could be
mounted directly to the same wheels. Therefore, the decision was made to use these new ‘Avid HS1’
rotors to improve the vehicle’s braking capacity at no additional cost.

4.5.2.1 Mounting Configurations

The type of mounting used for the brake calipers needed to be selected next. Two type of standard
mounting configurations (Figure 45) were analysed:

• Post Mount: calipers using this mounting

configuration are more readily available.
However, mounting holes would be
perpendicular to the main plate so an
additional bracket would be required.
• International Standard Mount: Less
Figure 45 – standard brake caliper mounting configurations
common configuration. However, the (Merlin Cycles,2014)
calipers could be face mounted directly
to the main plate.

Discussing with the Powertrain DMT and searching for potentially suitable calipers from the available
suppliers, it was found that post mount calipers were much more readily available in a range of prices,
dimensions and braking capacity. Therefore, the decision was made to design a bracket that would be
rigidly attached to the main plate and would provide appropriate mounting locations for post mount
calipers for 160 mm rotors.

The Avid technical drawings document (Avid, 2012)

proved to be an invaluable source of information for
determining the correct mounting points for the
required calipers. This information, summarised in
Figure 46, provided the correct location of the
caliper mounting points relative to the wheel axle.

Figure 46 – position of caliper mounting points relative to

wheel axle

DMT26| |p.56
DMT26 p.56
Using this information, a brake caliper arm was designed
that would be fastened to the main plate of the wheel
spindle subassembly, shown in Figure 47.

Initially, the idea was to mount the brake calipers towards

the top of the main plate, as shown in Figure 43a above,
Figure 47 – final brake caliper arm design
where they could be easily accessible if there was a need
to carry out maintenance of the brakes as well as to make sure that they would not interfere with the
other steering components. However, careful analysis of the geometric arrangement of the steering
components revealed that due to the caster angle of the wheels, the main plate would be at an angle
which would raise up the brake calipers, as shown on Figure 48.

This would have had the effect of increasing the

centre of gravity of the car due to the relatively
significant weight of the brake calipers and mounting
brackets, increasing the risk of rollover. Therefore, a
better idea was to mount the brake calipers towards
the bottom of the main plate. This is not normally
done on road vehicles due to packaging concerns
with the suspension components as well as the
position being more difficult to access, more
exposed to road debris and more prone to pad
Figure 48 – geometric analysis of steering components
damage due to flex of the disc under large braking
loads. For this application, the significantly smaller braking loads, the relatively clean track surface and
the enclosed design of the vehicle shell which largely protects against debris meant that it would be
possible to mount the brake calipers towards the bottom of the main plate.

4.5.3 Steering Column

The steering column consists of all the components that connect the steering interface of the vehicle
to the tie-rods that transfer the steering input to the front wheels. The exact mounting of the steering
column would depend on the available space towards the front of the chassis. However, the design
should ensure that the tie-rod which connects the front wheels to each other should be perpendicular
to the longitudinal axis of the vehicle when in its neutral (no steering input) position to ensure that
the Ackermann geometry is satisfied.

DMT26| |p.57
DMT26 p.57
From the decision matrix previously
analysed, it was decided that a handlebar
would be used as the steering interface for
the driver. This handlebar would have to be
rigidly attached to the chassis at least at
one location for the driver to comfortable
apply the required steering torque. The
front of the car was very limited for space
Figure 49 – use of universal joint with handlebar for steering interface
but the most suitable mounting point was
determined to be under the driver’s legs, such that the steering column would go between their legs
towards the floor of the vehicle. This was found to be the only option that would not restrict the
driver’s visibility but it meant that a mechanism would be required to control the angle between the
steering column and the floor of the car where it would be mounted. The best method to solve this
was found to be a universal joint that would allow a maximum angle of 45° of elevation for the
steering column, as shown in Figure 49, similar to the design used for the previous generation vehicle,
which was deemed to be sufficient for the driver to have an ergonomically acceptable driving
position.

The connection necessary between the steering column and the tie-rods was chosen to be a single
shaft, a Pitman shaft, that would be rigidly connected at one end to the universal joint previously
mentioned and at the other end to a lever referred to as a Pitman arm, as shown in Figure 50.

Figure 50 – Pitman connection mechanism between steering column and tie-rods

4.5.3.1 Ergonomics Consideration

Using a steering angle at the wheel of 15°, appropriate for a 6 m turning radius for a car with a
wheelbase of 1.55 m, it was possible to determine the relative motion of the Pitman arm and, hence,
the handlebars to decide if the operation of the steering system would be ergonomic. The Ackermann
steering geometry caused the tie-rod pivot connection at the wheel spindle assembly to trace a 15°

DMT26| |p.58
DMT26 p.58
arc of 37.5 mm radius while turning, corresponding to a horizontal distance travelled by the tie-rod
joint of the Pitman arm of approximately 10 mm.

An initial estimate for the length of the Pitman arm which allowed enough clearance for mounting all
other components was 30 mm in length, from the axis of rotation to the tie-rod mounting point, as
shown in Figure 51.

A horizontal distance of 10 mm travelled by the arm would have

resulted in an angular rotation of 19.5° of the Pitman arm. Since
the Pitman arm was to be rigidly connected to the steering
column through the use of a universal joint, this meant that the
handlebars at the other end of the steering column would also
have to rotate 19.5°. Initially this was found to be acceptable as
the driver could comfortably carry out this operation in the
limited space of the driver’s cockpit. However, increasing the
Pitman arm length to 37.5 mm allowed for an improved steering
ratio (defined as the ratio between the wheel angle and the
steering input angle at the handlebars) of 1:1. This is the same

Figure 51 – Pitman arm length steering ratio that is found in bicycles resulting in a familiar
steering sensation for the driver with the added benefit of
reducing the maximum rotation required at the handlebars from 19.5° to only 15°.

4.5.4 Stress Analysis of Steering Column Assembly

The design of the Pitman arm and Pitman shaft were heavily influenced by the loads experienced by
the steering system. It was expected that the loads acting on the Pitman shaft would be due to the
lateral cornering forces transferred to the Pitman arm through the tie-rods from the wheels. Axial
loads exerted by the driver on the steering system were assumed to be negligible.

4.5.4.1 Expected Loads

From previous analysis, the maximum lateral force occurring during cornering was calculated to be
323.73 N per wheel. This lateral force would act at the centre of the contact patch between the wheel
and road surface which was calculated to be 17.7 mm ahead of the kingpin axis due to the mechanical
trail generated by the 4° caster angle and the 254 mm radius of the wheels used (see ‘Introduction to
Adjustable Steering Angles’). The lateral force would try to self-centre the wheel by creating a
moment about the kingpin axis. Therefore, a force transferred through the tie-rod, 𝐹, is required to

DMT26| |p.59
DMT26 p.59
counteract this moment and to allow
the wheel to maintain its course around
a corner, as shown in Figure 52.

This tie-rod force would be transferred

to the wheel through the pivot point in
the Ackermann arm that is located 37.5
mm in front of the kingpin axis.
Therefore, a moment balance made it
Figure 52 – counteracting effect of lateral force
possible to calculate the tie-rod axial
load contribution, 𝐹, from each wheel.

𝐹 × 0.0375 = 323.23 × 0.0177

∴ 𝐹 = 152.7 𝑁

Applying a safety factor of 1.2, each wheel would require an axial tie-rod load of 184 N, resulting in a
maximum force of 368 N acting at the tie-rod pivot point of the Pitman arm.

A result of this analysis was the decision to mount the Pitman arm on the Pitman shaft as close as
possible to the nearest supporting bearing to minimise bending moments through the Pitman shaft.
However, a more significant consideration was the torsional deflection of the shaft.

4.5.4.2 Torsional Deflection of Pitman Shaft

The torsional deflection of a shaft is characterised by the Equation 20, where 𝜃 represents the angular
deflection in radians, 𝑇 is the applied torque, 𝐿 is the length of the shaft, 𝐺 is the shear modulus of
the material of the shaft and 𝐽 is the second polar moment of area for a shaft of diameter, 𝐷.

𝑇𝐿
𝜃= (20)
𝐺𝐽

𝜋𝐷 4
𝑤ℎ𝑒𝑟𝑒 𝐽 = (21)
32

The length of the Pitman shaft was expected to have an upper bound of 100 mm in order to minimise
space and total weight. Similarly, aluminium (6082-T6 grade with a shear modulus, 𝐺, of 26 GPa) was
initially considered due to its lightweight properties compared to steel. The total moment exerted on
the Pitman shaft was calculated from the length of the Pitman arm as well as the tie-rod axial forces
calculated previously.

𝑇 = 368 × 0.0375 = 13.8 𝑁𝑚

DMT26| |p.60
DMT26 p.60
Using these values and Equations 20 and 21, it was possible to calculate the minimum required
second polar moment of area and, therefore, the minimum required Pitman shaft diameter that
would result in a maximum acceptable torsional deflection taken to be 0.5°.

𝑇𝐿 13.8 × 0.1
𝐽= = = 6.08 × 10−9 𝑚4
𝐺𝜃 26 × 109 × 8.7 × 10−3

𝜋𝐷 4
∴ = 6.08 × 10−9
32

∴ 𝐷 = 0.016 𝑚 ⇒ 16 𝑚𝑚

This result suggests that the minimum diameter of the Pitman shaft was required to be 16 mm if it
was to be made from the aluminium material available. The same calculation was repeated for mild
steel, with a shear modulus, 𝐺, of 79.6 GPa, resulting in a minimum required diameter of 12 mm.

Although the required steel shaft would be heavier than the required aluminium shaft, the smaller
diameter meant that smaller bearings, bearing housings, universal joints and bolts could be used that
would decrease the overall weight and space required of the steel shaft option compared to the
aluminium shaft alternative. Therefore, the mild steel Pitman shaft option was selected.

4.5.4.3 Mounting of Pitman Arm on Pitman Shaft

The Pitman arm could be secured in place on the Pitman shaft through the use of a set screw, a key
and corresponding keyway or from the friction provided by a nut screwed into the Pitman shaft. The
latter option would be the most convenient as it would not require any additional features to be
designed into the Pitman arm or shaft apart from an external thread on one end of the Pitman shaft.
The only important consideration was to ensure that the nut could be tightened hard enough so that
the 13.8 Nm torque generated by the Pitman arm would not cause it to come loose.

Table 29 shows the average recommended tightening torques for readily available Grade 8.8 steel
nuts compiled from different fastener suppliers, including Grampian Fasteners (2017), Fastener Mart
(2015) and William Tools Co. (2019).

Table 29 - tightening torques for Grade 8.8 steel nuts

Standard Nut Size Average Recommended Tightening Torque/ Nm
M5 7
M6 11.8
M8 28.8
M10 57.3
M12 99.8

DMT26| |p.61
DMT26 p.61
It was found that M8 nuts or larger could be tightened during assembly to a torque greater than that
exerted on the nut by the Pitman arm. Ultimately, since this was a critical component for the safe
operation of the vehicle, an M10 nut was used to secure the Pitman arm onto the Pitman shaft.

4.5.4.4 Stress Calculations of Pitman Arm

Another design consideration was the geometry of the Pitman arm to ensure that it would not yield
under the maximum tie-rod loads described previously. As shown in Figure 50 above, the Pitman arm
is a lever that can be considered as a cantilever beam built-in at the point where it is connected to the
Pitman shaft. The maximum bending moment experienced by this component would be 13.8 Nm as
derived previously. An initial estimate for the thickness of this component was 4 mm as this was
successfully implemented in other components in the steering system. Similarly, an initial breadth of
11 mm was used to minimise the total weight of the components. Using Equation 6 as before, the
maximum bending stress for this geometry was found to be 171 MPa, as shown below, which was too
high compared to the yield strength of aluminium.

𝑀𝑦 13.8(0.0055)
𝜎= = = 171 𝑀𝑃𝑎
𝐼 0.004 × 0.0113
( 12 )

The calculation was repeated using a breadth of 16 mm resulting in a maximum bending stress of 80.9
MPa. This was acceptable and made it possible to manufacture this component from aluminium
instead of steel without increasing the size significantly, reducing the weight of this component.

4.5.4.5 Pitman Shaft Bearing Selection

A decision required for the design of the steering column components was to select appropriate
bearings for the Pitman shaft. As mentioned previously, the axial loads were taken to be negligible
and the maximum radial load would be 368 N. At least two bearings would be required to support the
shaft which is why pillow block bearings, also referred to as pedestal bearings, were preferred as they
would not need separate bearing housings to be designed and manufactured and they could be
installed directly on the floor of the vehicle. The majority of metal pedestal bearings available from
approved suppliers were found to be very heavy (approximately 500 grams each) and expensive
(approximately £20 each). Therefore, polymer bearings were investigated, paying careful attention to
their load ratings.

A much more lightweight and affordable solution was found in the ‘IGUS KSTM-12 mm’ pedestal
bearings, made from impact resistant and rigid thermoplastic composite material. The maximum
static radial load rating was found to be 2200 N for short term operation and 1100 N for long
operation, both adequate for the required use. Therefore, two of these pedestal bearings were

DMT26| |p.62
DMT26 p.62
chosen to be mounted directly on the floor panels of the chassis to support the Pitman shaft and the
steering column components.

4.5.4.6 Tie-rod Selection

Lastly, an appropriate material needed to be selected for the tie-rods, which would be loaded in
tension and compression only due to their pin-jointed structure, to ensure that they would not fail by
buckling. Therefore, it was important to compare the axial loads acting on the tie-rods to the
minimum force required for buckling, 𝑃𝑐 , shown in Euler’s buckling formula for a pin jointed truss.

𝜋 2 𝐸𝐼
𝑃𝑐 = (22)
𝐿2

To reduce weight, the team decided to try to implement the smallest available diameter of roll
wrapped carbon fibre tubes due to their high strength to mass ratio. The smallest available tubes
from Easy Composites Ltd, an approved supplier for this project, had an outer diameter of 10 mm and
an inner diameter of 8 mm. The flexural rigidity for these tubes, corresponding to 𝐸𝐼, was obtained
directly from the supplier (Easy Composites Ltd, 2019) and found to be 30 Nm2. After making an initial
model of the steering assembly, the distance from the Pitman arm tie-rod joint on the steering
column assembly to the Ackerman arm tie-rod joint on each wheel was 172.4 mm. This information
was then used to determine the critical load for buckling from Equation 22.

𝜋 2 × 30
𝑃𝑐 = = 9961 𝑁
0.17242

This critical load for buckling is significantly higher than the expected 184 N of maximum axial load
that will be experienced by the tie-rods. Therefore, it was determined that buckling would not be a
potential failure mode for this component. However, it was also necessary to investigate potential
failure by tensile or compressive yield.

The axial ultimate tensile stress of the 10 mm outer diameter and 8 mm internal diameter carbon
fibre tubes was also obtained from the supplier (Easy Composites Ltd, 2019) and found to be 600 MPa
in tension or 570 MPa in compression. The geometry of these chosen tubes resulted in a cross-
sectional area of 2.83x10-5 m2. The maximum axial stress on the tie-rods was then calculated as
follows.

𝐹𝑜𝑟𝑐𝑒 184
𝜎𝑎𝑥𝑖𝑎𝑙 = = = 6.51 𝑀𝑃𝑎 (23)
𝐴𝑟𝑒𝑎 2.83 × 10−5

The maximum axial stress was found to be significantly lower than the reported tensile or
compressive ultimate tensile stress confirming that these would also not be potential modes of
failure. Therefore, the selected carbon fibre tubes were deemed acceptable for this application.

DMT26| |p.63
DMT26 p.63
 4.5.5 Mounting of Steering System to Chassis
The steering column assembly would be connected to the floor of the chassis, approximately below
the driver’s knees using M5 bolts, nuts and washers on the two pillow block bearings that would
constrain the Pitman shaft assembly and steering column. This would only allow the column to be
tilted upwards up to an angle of 45° due to the universal joint, to allow the driver to comfortably get
into and out of the car and to be able to steer the vehicle in a more ergonomic position.

The wheel spindle assemblies were a more difficult challenge because they involved connecting the
‘L’ brackets mentioned previously to the round chassis tubes in a way that also provided sufficient
angular adjustment to change the camber setting of the vehicle.

An option was studied and developed which consisted of using a pair of two-piece shaft collars that
could be tightened around the appropriate tubes and would provide mounting points for plates that
would have slots cut out in order to provide the required angular movement for the camber
adjustment of the ‘L’ bracket and wheel spindle. The chosen configuration is shown in Figure 53.

The chosen collars were ‘RULAND OF-MSP-20-A’ which were

made of aluminium in order to keep the total weight to a
minimum, compared to the other option which used mild steel.
These collars were very suitable for this application because they
had flat sides and M6 tapped holes already machined into each
side, providing an ideal mounting surface for the chassis plates
with the camber slots. This option was much more expensive than

Figure 53 – steering to chassis mounting the adhesive alternative or other simpler shaft collars available.
mechanism concept
However, other more affordable options would have required
the team to machine flats and tapped holes on the circular surfaces of the collars which would have
required significantly more manufacturing time. The added cost of this option was justified.

The only remaining task involved the design of the plates which would include the slots for the
camber adjustment of the wheels. The most important design consideration was to ensure that the
slots were long enough to allow for a ±3° adjustment of the camber setting, as required by the
product design specification. The remaining dimensions of these plates and positioning of the slots
were selected to ensure that the adjustment in camber setting would result in no change to the track
width which is measured from the centre of the contact patches of the front tyres, according to the
official 2020 rules (Simons, 2019). The thickness of the plates was also set to 4 mm as this resulted in
a maximum bending stress of 28 MPa on each plate due to the maximum braking forces which was
acceptable for the grade of aluminium being used (6082-T6 grade).

DMT26| |p.64
DMT26 p.64
4.6 Final Steering Assembly Design
Very slight modifications were made to the provisional design to satisfy manufacturing considerations
and to increase the ergonomics of the driver inside the car:

• The main plate was modified slightly to include diagonal supports between the longer parts
that extended past the central section of the plate, which is where the brake and Ackerman
arms were connected, for added rigidity.
• The location of the mounting points for the brake and Ackerman arms were modified so that
the Ackerman arm and tie-rods would now be facing the front of the vehicle. This allowed the
steering column to be mounted a few centimetres forward to increase the space in the
driver’s cockpit.
• The tie-rod configuration in the steering system was modified to include a single drag-link that
extends from the Pitman arm to the front right wheel, with a tie-rod then connecting the right
wheel to the left wheel directly. This configuration allowed the Ackerman steering geometry to
be preserved for all steering inputs.

The final steering assembly is shown in Figure 54.

Figure 54 – final steering system design

DMT26| |p.65
DMT26 p.65
4.7 Manufacturing and Assembly of Steering System
The majority of the components required for the steering assembly were to be manufactured from
6082-T6 aluminium that could be sourced directly from the ME stores. The fasteners required for the
assembly of this system were also sourced from the ME Stores and were made from mild steel. Other
components that would be purchased from external suppliers were the rod-ends for the ‘L’ bracket
subassemblies as well as the carbon fibre tubes used for the drag-link and tie-rod components. These
were purchased from RS Components and Easy Composites Ltd, respectively.

4.7.1 Manufacturing Methods of Components

A discussion with Paul Woodward and Neil Beadle, responsible for fulfilling laser cutting and CNC
requests, respectively, revealed that the number of parts that each DMT team could request to be
manufactured by these methods would be very limited. Therefore, these were kept to a minimum.

The final list of steering parts that were requested to be laser cut were as follows:

Table 30 - steering parts for laser cutting

Part Name Part Description Quantity
ST-MAIN-PLATE-01-P Main plate of wheel spindle 2
subassembly.
ST-PITMAN-ARM-01-P Pitman arm where drag-link 1
connects to steering column.
ST-CHASSIS-PLATES-01-P Plates which connect the steering 4
system to chassis, providing
camber adjustment.

Similarly, the final list of steering components that were requested to be manufactured on the CNC
milling machine were as follows:

Table 31 - steering parts for CNC milling

Part Name Part Description Quantity
ST-ANGLED-BRACKET-BOTTOM- Bottom angled bracket of wheel 2
01-P spindle subassembly.
ST-ANGLED-BRACKET-TOP-01-P Top angled bracket of wheel 2
spindle subassembly.
ST-BRAKE-ARM-01-P Brake caliper mounting arm. 2

The angled brackets were required to have an angle of 102° between the inner faces to satisfy the
kingpin axis inclination requirement described in the design of these parts. This would be difficult to
produce accurately by manual machining which is why they were included in the CNC list. Similarly,
the brake caliper mounting arms required a complex shape to allow for the mounting of a wide range
of standard bicycle brake calipers which is why they were also submitted as a CNC milling request.

DMT26| |p.66
DMT26 p.66
The laser cutting process was a much quicker process which is why the team decided to use it as a
method for manufacturing all planar parts of the steering system including the chassis mounting
plates which featured camber slots for which the tolerance of the laser cutting process (±0.2 mm)
was deemed acceptable.

The remaining components of the steering system, shown in Table 30, were manufactured manually.

Table 32 - manually manufactured steering components

Part Name Part Description Quantity
ST-ACKERMANN-ARM-01-P Mounting point on wheel spindles 2
for tie-rod.
ST-KINGPIN-SPACER-LONG-01-P Long kingpin axis spacer. 2
ST-KINGPIN-SPACER-SHORT-01-P Short kingpin axis spacer. 4
ST-L-BRACKET-LEFT-01-P Left ‘L’ bracket. 1
ST-L-BRACKET-RIGHT-01-P Right ‘L’ bracket. 1
ST-M5-STUDDING-130-01-P 130 mm M5 threaded rod, acting 2
as kingpin axis.
ST-PITMAN-SHAFT-01-P Pitman shaft of steering column. 1
ST-RODEND-SPACER-3-01-P 3 mm rod-end spacers. 11
ST-SHAFT-01-P Main front wheel shafts. 2
ST-SHAFT-SPACER-01-P Shaft spacers used at either end of 4
each wheel shaft.
ST-TIEROD-INSERT-LEFT-01-P Tie-rod insert with internal left- 2
hand M5 thread.
ST-TIEROD-INSERT-RIGHT-01-P Tie-rod insert with internal 2
standard M5 thread.
ST-DRAGLINK-SHAFT-01-P Carbon fibre drag-link tube. 1
ST-TIEROD-SHAFT-01-P Carbon fibre tie-rod tube. 1

4.7.2 Manufacturing Progress

A manufacturing plan was implemented to ensure that responsibilities were distributed evenly among
all team members, taking advantage of the individual strengths of each person regarding their
previous workshop experience. This manufacturing schedule, which can be found in the Project
Quality Plan document of this project (Al-Shabazz, et al., 2020), made it possible to complete the
manufacturing of the steering system well within the available time.

The components shown in Table 30 required mostly manual turning, while others required manual
milling. Similarly, several of the parts that were submitted to be manufactured by CNC milling were
returned with a number of operations still remaining, such as the manual filleting of sharp corners
and the drilling of the required mounting holes using the milling machine shown in Figure 55.

DMT26| |p.67
DMT26 p.67
 Figure 55 – manufactured steering components

Remaining components such as the kingpin axis spacers, the wheel shaft spacers, the Ackermann
arms and the Pitman shaft were also manufactured manually due to their simple shape which allowed
the wheel spindle subassemblies to be completed, as shown in Figure 56.

Figure 56 – wheel spindle subassemblies

Some components of the steering assembly, such as the Pitman shaft and the ‘L’ brackets, required a
small amount of time more than was initially anticipated. However, other components, such as the
drag-link and tie-rod carbon fibre tubes and their inserts were manufactured without significant
setbacks. Overall, the distribution of manufacturing jobs across all members of the team allowed the
steering assembly to be completed in the time expected.

4.7.3 Mounting of Steering Assembly onto the Chassis

After the bonding of the chassis carbon fibre tubes was completed, the shaft collars that were used to
mount components onto the chassis were fitted around the tubes at their required locations. To
mount the steering system to the chassis it was necessary to securely attach two of these collars to
each side of the car at the front vertical tubes of the chassis.

DMT26| |p.68
DMT26 p.68
The collars were chosen to have an internal diameter of 20 mm to
match the external diameter of the tubes. However, the location of
these collars was required to be very close to the joints where
additional layers of carbon fibre were added in the wet lay-up process
to join several tubes together. Figure 57 shows the increased diameter
of the tubes where these collars needed to be mounted.

Significant manual filing was required, following the required health and
safety consideration to minimise exposure to carbon fibre dust, to
reduce the effective diameter of the tubes at those locations and to Figure 57 – effect of lay-up on
member diameter
allow the collars to clamp around the tubes as intended. Particular
attention was given to the amount of filing necessary to minimise the number of fibres that would be
compromised by the cutting action of the filing process.

The mounting process of the steering system onto the chassis was then continued, leaving only the
steering column subassembly to be incorporated. The two front wheel spindle assemblies were then
connected together using the tie-rod. Following this, the drag-link was connected to the front right
wheel spindle assembly at one end, with the other end being connected to the Pitman arm that had
been already fastened to the Pitman shaft. The last step of the process involved fastening the two
pedestal bearings which support and constrain the Pitman shaft onto the floor of the vehicle.

It was observed that the tie-rod would collide with the drag-link and the Pitman arm at the most
extreme steering input angles. To solve this, a riser was prototyped to elevate the Pitman shaft and
pedestal bearing assembly to prevent interference with the tie-rod. This marked the completion of
the design, manufacture and assembly phases of the steering system which is shown in Figure 58
mounted to the chassis with the front wheels removed.

Figure 58 – complete assembly of steering system

DMT26| |p.69
DMT26 p.69
5. TESTING AND MODIFICATIONS
One of the most important parts of this project is the testing of all the components that have been
designed and manufactured. The main purpose of this process is to corroborate the proposed designs
to determine whether they contribute to the main aims and goals of the project in the way that was
intended. Additionally, the testing phase provides a final opportunity to observe which components
could be modified to enhance the functionality of the overall product, an example of which was the
steering column riser that had to be implemented.

An extensive testing phase for the entire vehicle had been formulated in the Product Quality Plan for
this project (Al-Shabazz et al., 2020a). However, this plan was abruptly halted due to the lockdown,
forcing the team to suspend all further testing. Nonetheless, a significant amount of very valuable
information was collected through limited testing in the remaining days of the project.

5.1 Components Inspection

One of the methods in which the vehicle was tested involved setting the vehicle on the ground and
allowing a 60 kg person to climb onto the car, holding all three wheels to prevent the vehicle from
moving. With the test subject sitting on the car, steering linkages and components were inspected.

The chassis was initially set up with zero toe angle as well as with zero camber angle. However, under
the load exerted by the driver’s weight, the camber angle was found to change to 2° of negative
camber. This was a very significant change that would have observable effects on the handling
characteristics as well as reducing the overall efficiency of the vehicle. This, however, was not an
unexpected observation. It was known from the beginning of the project that the required lightweight
nature of all the components of the vehicle, including the chassis, would inevitably result in a less stiff
structure overall that would potentially deflect under load. This was the primary reason to include a
method for adjusting the camber angle even though it was agreed that zero camber would be
beneficial under all conditions.

To correct for this change in camber angle under load, the team proceeded to pre-set the camber to
2° of positive camber, when unloaded, so that the loads exerted by the driver would then correct the
angle to the required 0°. However, it is important to note that this adjustment is dependent on the
exact weight of the driver as well as the weight distribution of the vehicle after all components have
been mounted, including the powertrain parts and the shell. Therefore, the set-up configuration
would have to be repeated once all of these factors have been finalised.

DMT26| |p.70
DMT26 p.70
Another area of improvement that was discovered
during the testing procedure of the steering
components was the chosen universal joint for the
steering column assembly. This part was found to
break easily once the driver applied minimal effort
on the handlebars to try to turn the front wheels, as
shown in Figure 59.

This failure was also not unexpected due to the

plastic construction of the universal joint. However,
Figure 59 – failure mode of universal joint
if this part would fail, it was expected to do so under
the dynamic conditions of the competition, were the torque exerted on the universal joint could
reach values up to 13.8 Nm which would be higher than the rated strength of the plastic casing. The
team then proceeded to select a replacement steel universal joint from RS Components that would
satisfy the strength requirement as well as the diameter requirements determined by the other parts
that had already been manufactured, such as the Pitman shaft.

This universal joint, however, could not be purchased and implemented due to the abrupt ending of
the final stages of the project due to the COVID-19 situation. It is for this reason that the team
refurbished the broken component and mounted it on the steering column system to serve as an
indication of what the final assembly would have looked like.

5.2 Ergonomics Testing

An important design consideration that became a priority during the design phase of this vehicle was
the driver ergonomics. This is particularly important in this type of endurance competition were the
driver will be in an enclosed cockpit for an extended period of time. Therefore, every effort was made
to ensure that the driver would be as comfortable as possible, not only sitting inside the car but also
while operating the steering interface that was selected for this vehicle.

A valuable piece of feedback received from the test driver sitting in the car was that the steering
column was too short, making it uncomfortable to lie back on the seat of the car while at the same
time trying to reach for the handlebars, as shown in Figure 60 below. The team decided that the
steering column needed to be increased in length from the original value of 350 mm. However, the
length could not be extended too much because that would bring the handlebars too close to the
driver’s chest causing their elbows to uncomfortably extend outwards, potentially colliding with the
vehicle’s chassis or shell while turning. Therefore, the decision was made to extend the column by

DMT26| |p.71
DMT26 p.71
200 mm. It had been planned to implement this by purchasing a replacement carbon fibre tube of the
correct dimensions, although this could not be completed due to the COVID-19 situation.

Figure 60 – ergonomic testing

5.3 Regulation Compliance

One of the technical requirements for the Shell Eco-marathon Prototype Class of vehicles is that they
must have a turning radius of 8 metres or less (Simmons, 2019). This requirement would be tested at
the competition by making the car follow a 90° arc of a circle and ensuring that the radius described
by the car’s motion is 8 metres or less. The process would be repeated for both turning directions.

To ensure this vehicle satisfied this specification, the car was taken
to the service road directly outside the Pit Garage, where a
semicircle of 8 metre radius was marked on the ground using
masking tape.

The test procedure used to ensure that the car would pass the same
test at the competition began by aligning the outside tyre with the
semicircle markings so that the wheel was tangential to this
semicircle, with the steering being set to full-right lock, as shown in
Figure 61.
Figure 61 – turning radius test
The car was then manually pushed forward, while holding the full-
right lock on the steering system, until the external wheel was at 90° relative to its starting position,
as demonstrated by Figure 62. This would mean that the car had travelled along a 90° arc as required
by the rules which would then allow the turning radius to be measured.

DMT26| |p.72
DMT26 p.72
 Figure 62 – turning radius test results schematic

The tests were carried out twice in each direction to ensure reliability of the results and it was found
that the turning radius to the right was 6.35 m while the turning radius to the left was 5.15 m. The
difference in turning radius for each direction was believed to originate from small differences in
wheel alignment, due to the individual camber adjustment of each wheel. A small misalignment
could, therefore, result in very noticeable differences in turning radius, as was the case in this test.
Nonetheless, both turning radii recorded passed the required turning radius test comfortably.

This test was carried out without a driver sitting in the car in order to mitigate all potential risks
associated with the test and to be able to obtain a value that could be used for reference. However, a
more accurate result would have been obtained with all other systems mounted on the car, including
the powertrain components and the shell, with the designated driver sitting in the car and with the
wheels correctly adjusted to minimise the turning radii difference due to turning direction.

5.4 Additional Planned Testing

A number of additional tests had been planned, not only to ensure that the steering system would be
rule compliant, but also to maximise the efficiency of the vehicle during race. However, these tests
could not be carried out due to the COVID-19 disease outbreak.

The regulations require the steering system to be designed in such a way that the wheels would never
come into contact with any other components of the car such as the chassis, the shell or the wheel
bulkheads mandated by the rules (Simmons, 2019). This was due to be investigated after the shell had
been mounted onto the chassis and with a driver sitting inside the car who could provide visual
feedback of the results.

The effectiveness of the M5 threaded rod acting as the kingpin axis of the steering system was also
going to be investigated to ensure that it could withstand the dynamic loads exerted on the front

DMT26| |p.73
DMT26 p.73
wheels during competition conditions. This test could only be performed once the motor and battery
were installed on the vehicle and the procedure would consist of inspecting those components before
and after a test run, looking for noticeable permanent deflection or failure of the threaded rod. A
simpler test was also planned to be performed by applying a relatively high number of cycles of static
loads, simulating the driver getting into and out of the car. The limited amount of testing that was
carried out showed no indications that this component could cause any problems, even after having
several test drivers of above 60 kg sit in the car.

A more qualitative test was also planned to evaluate the entire steering assembly as a whole. One of
the most important characteristics of a steering system is the feedback it provides to the driver about
road conditions as well as the transfer of loads and the vehicle dynamics of the car. This is more
difficult to measure and depends on a number of factors ranging from the angles chosen in the initial
set-up of the vehicle to the rigidity of the components used. An important test that had been planned
was for a test driver to follow a simple test track layout that would allow them to get a feel for the
accuracy of the steering system as well as to relay other important feedback including any observed
deflections, vibrations or the general responsiveness of the car. This test would also evaluate the
ability of the car to self-centre the wheels on the exit of a corner and the steering effort that would
need to be applied at the handlebars to be able to turn the car.

The effectiveness of the braking system also needed to be investigated to ensure that the driver could
apply sufficient braking force to keep the car stationary at a 20 % incline, as mandated by the
regulations (Simmons, 2019). This would have been carried out towards the end of the testing phase
as it required the entire vehicle to be fully assembled, including the shell, the powertrain components
and the driver sitting in the vehicle to account for all the components that would have added mass to
the finished vehicle. The results of this test would have allowed a judgement to be made regarding
the efficacy of the brake disc diameters and brake calipers selected by the Powertrain DMT group.

A critical component requiring further testing was the aluminium plate that would separate the
driver’s cockpit and the powertrain compartment and which would act as the roll panel that would
protect the driver in the event of a rollover accident. This panel was designed to withstand a
minimum of 700 N of force applied in any direction along its top edge, as required by the regulations
(Simmons, 2019). Nonetheless, further testing had been planned using a fixture that would allow
loads greater than 700 N to be exerted on the plate to mitigate against any possible material defects
that could become a safety concern for the driver.

DMT26| |p.74
DMT26 p.74
6. BUDGET
The total budget for this project was £5000, of which £1000 was provided directly by the Department
of Mechanical Engineering and the remaining £4000 was accessed by the Shell Eco-Marathon team
principal from the Imperial Racing Green’s central funds.

In total, £1325 was spent which was in line with similar past projects. The biggest cost originated from
the lightweight carbon fibre materials used for the chassis as well as the safety equipment and the
shaft collars necessary for the secure mounting of components to the spaceframe. A full overview of
expenditures is given below.

Table 33 - expenditure overview

Area Supplier Description Cost (£)
15x1m Carbon fibre tubes, Carbon Fibre narrow fabric
plain weave 25mmx15m, 50mmx15m, Composite
Easy shrink tape, 50ml methacrylate adhesive
Composites Ltd 1kg EL2 epoxy laminating resin + AT30 hardener 430.09
Spaceframe Advanced
Hackspace ABS+ 3D printed Lugs (x20) 20.00
4mm thick aluminium sheets
D8,10,20 aluminium round bar, 40mm square
ME Stores aluminium bar, D16 mild steel round bar 16.98
Easy
Steering Composites Ltd D10 Roll wrapped carbon fibre tube (0.5m) 12.61
Nuts (M10, M5 LH thread, M5, M6)
Various socket head bolts
ME Stores M5x130 Studding 2.50
Ruland Collar (OF-MSP-20-A) (x18)
Pillow block bearings (IGUS KSTM12)
Fasteners/Fixtures RS Components M5 rod end (POS 5EC L), M5 rod end (POS 5EC) 472.90
Aluminium sheets (3mm & 4mm thick)
ME Stores 20mm square steel bar, D20 round steel bar 58.22
Easy
Chassis Composites Ltd CF Foam Core Panel 980x480, 480x480 6mm thick 171.50
Half mask respirator, gas vapour filter, mist filter, filter
Safety Equipment RS Components retainer 139.90
Total = £1325

DMT26| |p.75
DMT26 p.75
 7. DISCUSSION AND CONCLUSIONS

7.1 Potential Component Failure and Mitigation (AWP)

Table 34 - failure modes with corresponding mitigation methods
Failure Mode Description Mitigation
Shell The new chassis has been designed Minor interference can be mitigated by cutting the
interference to fit within the existing shell. problematic areas (especially the rim of the shell)
with new However, dimensions of shell CAD manually. In the unlikely event of major
chassis model may be slightly different to interference, carbon fibre wet layup can be used to
reality due to old age. modify the shape of the shell.
Incompatibility Mounting holes are a standard Minor interference can be mitigated by manual
between front dimension for ‘Post-mount’ brake filing of the aluminium arm. More severe
brake caliper calipers. However, caliper body interference (where filing would compromise the
and front dimensions are brand-specific so bracket strength) would require a new arm to be
brake caliper there could be interference manufactured by CNC. Alternatively, use 3D
mounting arm between the caliper body and the printing with carbon fibre infused filament, similar
mounting arm designed. to that used for the rear caliper arm.
Excessive The long and narrow carbon fibre Use a carbon fibre tube with a larger diameter to
torsional tube used for the steering column increase its polar moment of area and reduce
deflection of may exhibit excessive torsional torsional deflection. This would require a new 3D
steering deflection that can hinder the printed bracket to connect this tube to the
column accuracy of the steering system and handlebars, and an insert to adapt the tube to the
reduce driver confidence in the car current universal joint or purchase a bigger
universal joint.
Camber The camber adjustment relies on Apply high-friction shims between ‘L’ bracket and
adjustment the friction between the ‘l’ bracket chassis mounting plates. Anti-vibration pads could
may change and the chassis mounting plates. also be used to increase friction as well as damping.
during Vibrations during operation may Effect of vibrations could be mitigated further by
operation cause the camber angle to change. using nyloc nuts and anti-vibration washers.
Inspection is advised after every
run.
Tie-rod and Aluminium inserts were bonded at Drill a small M2 hole through the carbon fibre tubes
drag-link each end of the carbon fibre tie-rod and aluminium inserts and connect the two using a
insert bonding and drag-link tubes. Bonding spring pin. Alternatively, if this results in cracking of
failure strength depends on many external the tubes, replacement solid aluminium rods should
factors (e.g. temperature and be made with the threads tapped out at each end
humidity) so there is a possibility of to be used as a last-resort (albeit heavier)
these bonds failing. replacement.
M5 rod-end The low static and dynamic loads Replace M5 components with similar M6 variants, if
failure in ‘L’ allowed the rod-ends to be loaded necessary. Purchase M6 rod-ends, M6 threaded rod
Bracket in bending without any observable for the kingpin axis, M6 nuts and manufacture
subassembly deflection. However, this could be a kingpin axis spacers, all to be used as potential
potential source of failure. spares. If needed, the M5 holes in the angled
brackets and ‘L’ brackets can be drilled out to M6
and re-tapped where necessary by hand at the
competition.
M5 bolt failure The shock loads and lateral loads Replace M5 bolt with an M6 bolt. If necessary, the
in the axle translated to the axle mount axle mount component can be inspected and
mount component from the wheel may measure in its angular misalignment mitigation
component position and an equivalent continuous, single part

DMT26| |p.76
DMT26 p.76
 cause this bolt to fail due to high can be manufactures using CNC to replace the two-
shear stresses. part component.
Failure of one When repeatedly loaded with the All cured joints of the spaceframe should be
or more cured driver’s weight and when used for inspected regularly and physically tested using
joints of the extended periods or exposed to hot weights. A record of chassis deflections should be
spaceframe road temperatures, the wrapped kept to identify when a joint has been
carbon fibre joints may weaken and compromised. In this case, reapply the layup
deflect. process with more layers to create a new
undamaged structure, re-test before use.
Insufficient When designing the chassis, a 5’ 5” If the driver’s vision is obstructed in any way, a wet
visibility for human model was used to confirm layup process can be used to alter the shell in order
driver that the vision of the driver was not to ensure full visibility is restored. Positions of
affected in any way. If the driver, windows on the shell may also be altered if
when suited up to race, exceeds the necessary.
size of the model, their vision may
be obstructed.

7.2 Conclusions
The overall objectives for the project were as follows:
1. Design a chassis that will comfortably fit an average female driver, provide mounting points
for all components and be mounted to the existing aerodynamic shell with minimal
modifications.
2. Design a steering system which maximises efficiency and provides an adequate level of
feedback to the driver, allowing for adjustments to be made to the front wheel alignment.
3. Implement a powertrain interface on the developed chassis that will be compatible with the
developments of Powertrain DMT Group 27.

The final chassis successfully fit an average female driver, provided mounting points for all
components and achieved a weight reduction of 23% compared to last year’s entry with all
components attached. Though the powertrain system had not yet been produced by the Powertrain
DMT team at the time of the project ending, the chassis interface was fully compatible with its design.
While thorough testing did not occur due to the reduced timeline of the project, the validity of the
final product in meeting loading conditions and achieving a high efficiency was backed by in-depth
stress and deflection analysis and steering system design calculations. Additionally, the steering
system incorporated adjustable front wheel alignment, allowing the efficiency of the system to be
maximised regardless of the driver. Although the project was expected to consume most of the
budget for the entire vehicle, significantly less than half of the budget was used at £1325. This project,
in conjunction with the work produced by the Powertrain team, will place the SEM team for the
following year in a very good position to have a completed vehicle, ready to compete in the race.

DMT26| |p.77
DMT26 p.77
8. INDIVIDUAL CRITIQUES
Jessica Eichel

Overall, it has been a pleasure working together with this highly motivated and talented team to
exceed our project aims and objectives. The team’s approach of emphasising simple engineering
evaluations and calculations to inform practical decisions on manufacturing proved successful and
allowed completion of the entire SEM chassis and steering system on time and before the lockdown
was implemented. The project management approach worked particularly well, whereby one
member ensured the Gantt chart remained up to date and each member was formally assigned
weekly tasks on Microsoft teams.

My main roles within the team were contributing to the design development of the chassis including
materials selection and manufacturing methods as well as prototyping and testing of the spaceframe
design and layup. I was also responsible for the procurement and budgeting, providing a valuable part
in ensuring all deliverables were met on time. Being part of a larger SEM team provided a wider
insight from our supervisor in weekly meetings, as well as the Team Principal, the Powertrain Team,
and the formula student team. The main improvements, preventing the SEM vehicle from being fully
finished, was a lack of collaboration early on with the Powertrain Team and the Team Principal,
projects in future years should aim to improve this relationship.

Carlos Firgau
Initially, this project seemed very ambitious due to its wide scope, although I particularly liked being
able to develop unique ideas in different areas of the vehicle. At first, we did not have the
independence that we had expected – many of the decisions needed to be confirmed by the Shell
Eco-marathon team principal and chiefs. This was different from other DMT groups but it ultimately
had a positive effect on our learning experience, being more representative of a real-world
engineering environment. This was especially true as we often had to rely on information provided by
the Powertrain DMT to be able to proceed in our own design. More frequent meetings would have
been beneficial and are an area of improvement for future projects.

I particularly enjoyed the position of Team Lead where I could make the operational decisions that
shaped our final design, relying on the essential knowledge and contributions from the rest of my
team. This project allowed me to delve deeper into areas of my own interest in automotive design
such as vehicle dynamics and steering characteristics, and to increase my knowledge of the behaviour
of carbon fibre. Ultimately, I believe that my methodical approach and attention to detail contributed
positively to the robustness of the carbon fibre chassis and the effectiveness of the steering system.

DMT26| |p.78
DMT26 p.78
Asad Raja
Despite the project being cut short, the team was able to produce the chassis and steering in its
essentially complete form and conduct some product testing. This is a mark of the team’s consistently
proactive working. An early parameter of efficient working that was set by the Project Supervisor was
completing the design phase of the project in first term, having the formal Design Review before the
holidays, which was achieved. As the Team Secretary, I was responsible for the project structure,
producing a detailed Gantt chart with sufficient contingency and holding meetings at least once a
week to evaluate progress and make action plans. This effective project planning, which was one of
the strengths of the project, meant that internal and external deadlines were met. A dedication to
Shell Eco-marathon beyond the DMT project, innovation and thorough analysis and the agreeable
nature of all team members to foster a positive dynamic were other integral strengths.

The main weakness of the project was occasional ineffective communication or unequal distribution
of workload. Team members would sometimes be modest or selective when updating the team on
their contributions. This would lead to a lack of clarity on the process that was taken to come upon
certain decisions or the progress of overall tasks. Though occasional more formal and thorough
reviews would reveal the true amount and distribution of work, which the project process would then
be retrospectively adapted to, a more honest and open approach could have been taken throughout.

Rahman Al-Shabazz
Prior to starting, I had my concerns on the challenges we would experience over the course of the
project given its large scope. However, I was fortunate in being able to work alongside intellectual and
driven individuals and am proud of the final result. This has been the best organised group I have
worked with over the course of my academic career, and I believe this played a big part in our
success. Weekly, in-person meeting where actions tasks were set and the effective use of the
OneDrive system to collect all our work, ensured the team made consistent and efficient progress.
The main area for improvement would be the communication with the Powertrain Team. We failed to
set robust deadlines for certain designs which meant our team would occasionally be waiting for
dimensions or other information before we could proceed with our work.

My main role in the team was the overall development of spaceframe structure using CAD. As well as
conducting my own research and developing ideas, this required that I communicated constantly with
all members to ensure their respective contributions to the chassis’ design did not interfere with each
other. As well as this, I was responsible for exploring the risks involved when working with carbon
fibre and constructed the risk assessments needed before manufacture could begin – amongst other
smaller tasks. Overall, I feel we worked very well together and am pleased with the final product.

DMT26| |p.79
DMT26 p.79
9. REFERENCES
Al-Shabazz, R., Eichel, J., Firgau, C. & Raja A. (2020a) ergonomic driver posture and safety. Periodicum
Project Quality Plan. 2nd Edition. London, United Biologorum.
Kingdom: Imperial College London.
Lin, W. & Yoda, T. (2017) Classifications, Design
Al-Shabazz, R., Eichel, J., Firgau, C. & Raja A. (2020b) Loading, and Analysis Methods. Bridge Engineering.
Progress Report. 1st Edition. London, United Kingdom: p.1-30.
Imperial College London.
Merlin Cycles (2014) Buyers guide to suspension forks.
Avid (2012) OEM Product Technical Specifications. Available from:
Available from: https://www.merlincycles.com/blog/buyers-guide-to-
http://www.dirtfreak.co.jp/cycle/sram/service/avid/av suspension-forks/
id_technical_specifications_my13_updates.pdf
Milliken, W. & Milliken, D. (1995) Race Car Vehicle
Childs, P. (2019) Mechanical Design Engineering Dynamics. 5th Edition. Warrendale, Pennsylvania,
Handbook. 2nd Edition. United Kingdom: Elsevier. United States of America: Society of Automotive
Engineers, Inc.
Easy Composites Ltd (2019) Roll Wrapped Carbon
Fibre Tube. Mechanical Properties and Specification. Occupational Health & Safety (2009) Carbon fibre
Available from: composites – OHS information sheet. Melbourne,
https://www.easycomposites.co.uk/#!/cured-carbon- Australia: Monash University. Available from:
fibre-products/carbon-fibre-tube/roll-wrapped- https://www.monash.edu/ohs/info-docs/safety-
carbon-fibre-tube/carbon-fibre-tube-roll-wrapped- topics/chemical-management/carbon-fibre-
8mm.html composites-ohs-information-sheet
Fastenal (2009) Fastenal Engineering & Design Riley, W. & George, A. (2002) Design, Analysis and
Support. Bolted Joint Design. Winona, Minnesota, Testing of Formula SAE Chassis. Ithaca, New York,
United States of America. United States of America: Cornell University.
Fastener Mart (2015) Metric Bolts, Screws & Nuts Santing, J., Onder, C., Bernard, J., Isler, D., Kobler, P.,
Tightening Torques. Available from: Kolb, F., Weidmann, N. & Guzzella, L. (2007) World’s
https://www.fastenermart.com/files/metric_tighten_t Most Fuel Efficient Vehicle. Design and Development
orques.pdf of PAC CAR II. Zurich, Switzerland: vdf
Hochschulverlag AG.
Fekete, J.R. (2017) Carbon Fibre Reinforced Polymer.
Automotive Steels. Available from: Seward, D. (2014) Race Car Design. 1st Edition. United
https://www.sciencedirect.com/topics/engineering/ca Kingdom: Palgrave.
rbon-fibre-reinforced-polymer/pdf
Shell Eco-marathon. (2020) Shell Eco-marathon 2020
Grabarczyk, M. (2003) Center of body mass and the Programme Update. Available from:
evolution of female body shape. American Journal for https://www.shell.com/make-the-future/shell-
Human Biology. p.144-150 Volume 15 Issue 2. ecomarathon/for-participants/2020-programme-
update.html
Grampian Fasteners (2017) Torque Settings. Available
from: Singh, R. (2010) Structural Performance Analysis of
https://www.grampianfasteners.com/files/95b2c19b- Formula SAE Car. Patiala, Punjab, India: Chitkara
1d29-4624-abdf-0813df2db3ac/Torque_Settings.pdf Institute of Engineering and Technology
Guinness World Records (2018) Most fuel-efficient Simmons, S. (2019) 2020 Official Rules. Chapter 1.
vehicle (prototype). Available from: Shell Eco-Marathon.
https://www.guinnessworldrecords.com/world-
records/most-fuel-efficient- Steinberg, D. (2000) Vibration Analysis for Electronic
vehicle?fb_comment_id=874403022578406_1981249 Equipment. 3rd Edition. New York, United States of
211893776 America: Wiley.

King, J.E. (1989) Failure in composite materials. Metals Waterman, B. (2011) Design and Construction of a
and Materials 5. Available from: Space-frame Chassis. Perth, Australia: University of
http://publications.aston.ac.uk/id/eprint/23791/1/Fail Western Australia
ure_in_composite_materials.pdf
William Tools Co. (2019) Standard Bolt Tightening
Kovacevic, S., Vučinić, J., Kirin, S. & Pejnović, N. (2010) Torque. Available from:
Impact of anthropometric measurements on https://www.wtools.com.tw/STANDARD-BOLT-
TIGHTENING-TORQUE.shtml

DMT26| |p.80
DMT26 p.80