Design and Analysis of Formula Student Race Car Wheel Rims

University Degree in Mechanical Engineering
2021-2022
Bachelor Thesis
“DESIGN AND ANALYSIS OF

LOW-COST CFRP WHEEL RIMS FOR
A FORMULA STUDENT RACE CAR”
Pablo Aparicio Alonso
Francisco Javier Velasco López

Leganés, September 8 2022
This work is licensed under Creative Commons Attribution – Non Commercial –

Non Derivatives
SUMMARY
Weight reduction is one of the fundamental pillars of race car design, as it allows for
a nimbler, more reactive, and therefore faster car around a given circuit. Whilst reduction
of suspended and non-rotating masses is simpler, reducing the non-suspended, rotating
masses has a larger impact on the handling characteristics of the vehicle.
Carbon Fiber Reinforced Polymer (CFRP) wheel rims are an effective solution to this
problem, but commercially available products are few and costly, whilst most Formula
Student teams operate on tight budgets. In order to offer a viable, low-cost, CFRP rim that
low-budget teams can use for their own purposes, the design will be optimized for ease
of mold manufacturing and hand lay-up using commercially available pre-impregnated
CFRP and pressure molding. This a simple and cheap process capable of achieving ex-
cellent mechanical properties and complex geometries with a very low manufacturing
cost.
Keywords: CFRP, Rim, Wheel, Formula Student, Low Cost.
iii
DEDICATION.
To my family, for their endless love and support all throughout these years. Words
cannot describe just how much you mean to me.
To my friends, who were there when I needed them, even if I wasn’t.
To Formula UC3M, for hard lessons well learned, and good times too often forgotten.
To the entire MAD Formula Team, but especially, and in no particular order, Jaime,
David, Ana, Fer, Jorge, Miguel, Sergio, Pablo, Jorge, Diego. Without your hard work and
constant dedication, none of this would have been possible.
To the global Formula Student community, sponsors, teams, volunteers, and staff, who
year after year make this absolute madness possible and worth it.
To everyone who understands that "Because racecar" is more than enough justification
to do what we do.
v
CONTENTS
1. INTRODUCTION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1. Motivation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2. Objectives. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3. Structure. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.4. Tools and Software used. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2. SOCIOECONOMIC ENVIRONMENT: FORMULA STUDENT. . . . . . . . . . 4
2.1. Brief history of Formula Student competitions. . . . . . . . . . . . . . . . . . . 4
2.2. The 2022 Formula Student Germany rules and its derivative rulesets. . . . . . 6
2.3. Formula Student Events and Scoring. . . . . . . . . . . . . . . . . . . . . . . . 6
2.4. MAD Formula Team . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
3. STATE OF THE ART.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
3.1. Commercially-developed rims. . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
3.2. Student-developed rims as state of the art. . . . . . . . . . . . . . . . . . . . . . 18
3.3. Why student-developed rims are the state of the art in FS. . . . . . . . . . . . . 21
3.4. Regulatory Framework: Formula Student regulations pertaining wheel rims. . 22
4. DYNAMIC ANALYSIS OF THE IMPACT OF CFRP RIMS ON RACE CAR
PERFORMANCE. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
4.1. Basics of FS Dynamic disciplines. . . . . . . . . . . . . . . . . . . . . . . . . . 24
4.2. Basic dynamics of a race car. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
4.3. Lap Time Simulation. Introduction to the Key Efficiency Metric. . . . . . . . . 26
4.4. Project Budget: Impact and viability analysis. . . . . . . . . . . . . . . . . . . 31
5. DEVELOPMENT OF THE FEA MODEL . . . . . . . . . . . . . . . . . . . . . . 35
5.1. Brief introduction to the Finite Element Method and software. . . . . . . . . . 35
5.1.1. Mesh quality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
5.2. Loads and load cases. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
5.2.1. Maximum Braking. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
5.2.2. Maximum Cornering. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
5.2.3. Tyre Mounting. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
vii
5.2.4. Overpressure. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
5.3. Load application. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
5.3.1. Dynamic load application.. . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
5.3.2. Mounting load application. . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
5.4. Material Characterization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
5.5. Fatigue. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
6. DESIGN AND ANALYSIS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
6.1. Design process. CAD model. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
6.2. Spoke design analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
6.3. Two-part vs. Single-part rim. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
6.4. Final design and FEA model. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
6.5. Mold design and manufacturing. . . . . . . . . . . . . . . . . . . . . . . . . . . 61
6.6. Analysis and results. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
7. CONCLUSIONS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
7.1. Design conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
7.2. Viability of low cost manufacturing. . . . . . . . . . . . . . . . . . . . . . . . . 75
7.3. Environmental Impact.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
8. FUTURE DEVELOPMENTS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
BIBLIOGRAPHY. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
viii
LIST OF FIGURES
Figure 1.1 Early race cars, often stripped down versions of regular road cars,
courtesy of Classic Cars journal. . . . . . . . . . . . . . . . . . . . . . . 1
Figure 2.1 University of Texas-Arlington 1986 champion FSAE race car, cour-
tesy of the University of Texas-Arlington News Service photograph col-
lection. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
Figure 2.2 University of Birmingham 1998 race car on the first Formula Stu-
dent UK competition, courtesy of Racecar Engineering Magazine. . . . . 5
Figure 2.3 Scoring for the Manual Acceleration Event, FSG rules 2022. . . . 8
Figure 2.4 Layout for the Skidpad figure-eight circuit, FSG rules 2022. . . . . 8
Figure 2.5 Scoring for the Manual Skidpad Event, FSG rules 2022. . . . . . . 8
Figure 2.6 Scoring for the Autocross Event, FSG rules 2022. . . . . . . . . . 9
Figure 2.7 Scoring for the Endurance Event, FSG rules 2022. . . . . . . . . . 9
Figure 2.8 Scoring for the Efficiency Event, FSG rules 2022. . . . . . . . . . 10
Figure 2.9 Scoring for the Engineering Design Event, FSG rules 2022. . . . . 10
Figure 2.10 Scoring for the Cost and Manufacturing Event, FSG rules 2022. . . 11
Figure 2.11 Scoring for the Business Plan Presentation Event, FSG rules 2022. 11
Figure 2.12 Formula UC3M C3R16 in Montmeló, Catalonia, Formula Student
Spain 2016. Courtesy of MAD Formula Team. . . . . . . . . . . . . . . . 12
Figure 2.15 Formula UC3M C3R19 in Hockenheim, Baden-Württemberg, For-
mula Student Germany 2019. Courtesy of MAD Formula Team. . . . . . 14
Figure 2.16 MAD Formula Team MFT01 at the car rollout event, Leganés,
Madrid, 2021. Courtesy of MAD Formula Team. . . . . . . . . . . . . . 15
Figure 3.1 Koegnisegg rim manufacturing, from the THE DRIVE youtube
channel. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
Figure 3.2 BRAID StuRace carbon, from the BRAID Website. . . . . . . . . 18
x
Figure 3.3 TU Graz rims and cross section, courtesy of TU GRAZ Racing
Team. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
Figure 3.4 UAS Graz rims, courtesy of FSG Media Team. . . . . . . . . . . . 19
Figure 3.5 University of Akron rims. . . . . . . . . . . . . . . . . . . . . . . 20
Figure 3.6 University of Kansas rims. . . . . . . . . . . . . . . . . . . . . . 20
Figure 3.7 TU Delft rims, courtesy of FSG Media Team. . . . . . . . . . . . 21
Figure 3.8 Article T 2.5 of the 2022 Formula Student Germany Rules. . . . . 22
Figure 3.9 Article IN 2.1 of the 2022 Formula Student Germany rules. . . . . 22
Figure 4.1 Formula Student Germany race track, Formula Student Germany
handbook. Notice the abundance of tight corners and lack of straight
sections. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
Figure 4.2 Laptime vs. Weight, Powertrain efficiency sweep chart, 15 steps.
Lower is better. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
Figure 4.3 Laptime vs. Weight, Powertrain efficiency sweep chart, 20 steps.
Lower is better. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
Figure 4.4 Average speed vs. Weight, Powertrain efficiency sweep chart, 15
steps. Higher is better. . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
Figure 4.5 Average speed vs. Weight, Powertrain efficiency sweep chart, 20
steps. Higher is better. . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
Figure 4.6 Fuel consumption vs. Weight, Powertrain efficiency sweep chart,
15 steps. Lower is better. . . . . . . . . . . . . . . . . . . . . . . . . . . 30
Figure 4.7 Fuel consumption vs. Weight, Powertrain efficiency sweep chart,
20 steps. Lower is better. . . . . . . . . . . . . . . . . . . . . . . . . . . 30
Figure 4.8 Empty KEM tool, before data entry. . . . . . . . . . . . . . . . . 32
Figure 4.9 Balance point on the KEM metric and final budget. . . . . . . . . 33
Figure 4.10 Expected budget for the project, with and without considering
sponsor collaborations. . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
Figure 5.1 Mesh quality settings for the rim FEA model. . . . . . . . . . . . 37
Figure 5.2 Example of a high quality section of the mesh, on the change of
curvature for the rim spoke. . . . . . . . . . . . . . . . . . . . . . . . . . 37
Figure 5.3 Dynamic load application point on the rim, from the FEA model. . 42
Figure 5.4 Mounting load application on the edge of the rim, from the FEA
model. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
xi
Figure 5.5 Overpressure load application on the rim, from the FEA model. . . 44
Figure 5.6 Adjusted Stress vs. Strain curve for sample 1. . . . . . . . . . . . 46
Figure 6.1 Four-spoke rim variant. . . . . . . . . . . . . . . . . . . . . . . . 50

Figure 6.2 Five-spoke rim variant. . . . . . . . . . . . . . . . . . . . . . . . 50
Figure 6.3 Twelve-spoke rim variant. . . . . . . . . . . . . . . . . . . . . . . 50
Figure 6.4 Lateral load application on the 4-spoke rim variant. . . . . . . . . 52
Figure 6.5 Von Mises stresses, in MPa, for the 4-spoke rim variant submitted
to lateral load. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
Figure 6.6 Displacement, in mm, for the 4-spoke rim variant submitted to
lateral load. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
Figure 6.7 Longitudinal load application on the 4-spoke rim variant. . . . . . 54
Figure 6.8 Von Mises stresses, in MPa, for the 4-spoke rim variant submitted
to longitudinal load. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
Figure 6.9 Displacement, in mm, for the 4-spoke rim variant submitted to
longitudinal load. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
Figure 6.10 Preference in terms of overall aesthetics by 17 members of MAD
Formula Team. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
Figure 6.11 Parametric dimensions of the lip. . . . . . . . . . . . . . . . . . . 59
Figure 6.12 Final model of the CFRP rim. . . . . . . . . . . . . . . . . . . . . 59
Figure 6.13 Element size vs. Displacement (both in mm) for the mesh sensi-
tivity analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
Figure 6.14 Element size vs. Other relevant variables for the mesh sensitivity
analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
Figure 6.15 3D printing of one of the mold components. . . . . . . . . . . . . 62
Figure 6.16 Mold for the rim bucket, separated into its two components. . . . . 62
Figure 6.17 Mold for the rim bucket, with both components assembled together. 63
Figure 6.18 Mold for the rim bucket, joined with threaded rods before lamina-
tion. Small damage to the upper lip was easily repaired. . . . . . . . . . . 64
Figure 6.19 Molds for the rim spokes, separated. . . . . . . . . . . . . . . . . 65
Figure 6.20 Male (plunger) mold for the spokes. . . . . . . . . . . . . . . . . 65
Figure 6.21 Molds for the rim spokes, assembled together. . . . . . . . . . . . 66
xii
Figure 6.22 Final Von Mises stresses, in MPa, for the lateral acceleration load
case. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
Figure 6.23 Final Von Misses stresses, in MPa, for the longitudinal accelera-
tion load case. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
Figure 6.24 Final displacements, in mm, for the lateral acceleration load case. . 68
Figure 6.25 Final displacements, in mm, for the longitudinal acceleration load
case. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
Figure 6.26 Final composite ply failure index for the lateral acceleration load
case. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
Figure 6.27 Final composite ply failure index for the lateral acceleration load
case. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
Figure 6.28 Final weight of the model in tons, as per the FEA suite. . . . . . . 70
Figure 6.29 Final displacement, in mm, for the tyre mounting load case. . . . . 71
Figure 6.30 Final displacement, in mm, for the overpressure load case. . . . . . 71
Figure 6.31 Final Von Misses stresses, in MPa, for the tyre mounting load case. 72
Figure 6.32 Final Von Misses stresses, in MPa, for the overpressure load case. . 72
Figure 6.33 Final composite ply failure index for the tyre mounting load case. . 73
Figure 6.34 Final composite ply failure index for the overpressure load case. . 73
xiii
LIST OF TABLES
Table 4.1 Best times, corresponding score and points delta to the best score at
FSG2022 Autocross Event for different wheel rim configurations. . . . . 31
Table 5.1 Mechanical properties for CMCFRP, as per samples tested. . . . . . 46
Table 6.1 Maximum Von Mises stress and absolute displacement for each
spoke configuration and dynamic load case. . . . . . . . . . . . . . . . . 56
Table 6.2 Differences in stress and displacement for one- and two- part rims. . 58
Table 6.3 Von Mises Stresses, displacements and composite failure index for
each laminate and spoke thickness iteration, plus the weight for the cor-
responding rim. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
xv
xvi
NOMENCLATURE
CAD : Computer Aided Design
CFRP : Carbon Fiber Reinforced Polymer
CMCFRP : Compression Molded Carbon Fiber Reinforced Polymer
FEA/FEM : Finite Element Analysis/Finite Element Method
FS /FS AE : Formula Student/Formula Society of Automotive Engineers
FS G : Formula Student Germany
KEM : Key Efficiency Metric
LT S : Lap Time Simulation
MADFT : MAD Formula Team
MFT 01 : MAD Formula Team’s prototype race car for the 2021 season
PMD : Project Management Document
Prepreg : Pre-Impregnated Carbon Fiber
xvii
1. INTRODUCTION.
1.1. Motivation.
Race car design is a very specific discipline within the scope of automotive design, and
carries a significant shift to the priorities of vehicle design. Race cars make very few
concessions with regards to comfort and almost none with regards to practicality, and
whilst regular cars are concerned with a wealth of international regulations, requirements
for mass manufacturing, and market pressure to make sales, race cars are typically only
burdened by the specific rules for the competition they participate in, and, perhaps more
importantly, the budget that can be allocated to their construction.
From the very onset of automobile racing, it was established that a smaller car carrying
less weight, and therefore smaller inertias, was faster around a track than a larger, heavier
car. This is due to the fact that the smaller, lighter car is able to transition through the three
most important stages in racing, braking-cornering-acceleration, carrying more speed than
its heavier counterpart. As racing typically involves going around a circuit composed by
a series of connected corners, the differences in speed quickly add up and can easily
determine the victor, all else being equal.
Fig. 1.1. Early race cars, often stripped down versions of regular road cars, courtesy of Classic
Cars journal.
Broadly generalizing, we can speak of sprung mass, that which is supported by the
suspension system of the car, and non-sprung mass, that which is supported by the surface
of the road; and also of rotating mass and non-rotating mass, the difference between the
two hopefully being obvious. Although any reduction in mass is beneficial, it is generally
accepted that it is more effective to reduce the non-sprung mass versus the sprung mass,
and the rotating mass versus the non-rotating mass.
Optimizing the weight of every single component on a race car is a problem of enor-
mous complexity and beyond the scope of any single project. However, a much simpler
1
and thoroughly solvable problem can be found in the optimization of the weight of a
single, non-sprung, rotating element of the car, the wheel rim. By using lightweight
composite materials with excellent mechanical properties, such as Carbon Fiber Rein-
force Polymer (CFRP), a significant weight reduction versus metallic alloy rims can be
obtained.
1.2. Objectives.
Having outlined the main parameters of the problem: budgetary constraints, importance
of weight reduction, and benefits of reducing non-sprung rotating mass; as well as the
problem itself, we can now state the objectives of this project:
• Understand how the weight of the rim affects the dynamic capabilities of the car, its
handling and on-track performance on all Formula Student dynamic events.
• Understand the loads that the rims will have to withstand in order to be safe, usable
and efficient.
• Characterize the mechanical properties of short-strand Compression Moulded Car-

bon Fiber Reinforced Polymer (CMCFRP).
• Design and perform Finite Element Analysis (FEA) of a functional lightweight rim
using pre-impregnated Carbon Fiber Reinforced Polymer and short-strand CM-
CFRP, compatible with the 2022 Formula Student prototype developed by MAD
Formula Team.
• Adapt the design to facilitate low-budget manufacturing within the capabilities of

MAD Formula Team.
1.3. Structure.
The first part of this document is dedicated to the introduction and objectives for the
project, as well as the structure of the document itself and the tools used for the develop-
ment of this work.
Part two provides an overview of the framework for the Formula Student competitions,
which will include a brief introduction to the competition itself, its regulations, and its
history, as well as an introduction to the Formula Student team at the University Carlos
III de Madrid.
The third part of the document will cover the state of the art in terms of CFRP rims,
both in and out of Formula Student, including both commercial alternatives by reputed
manufacturers and self-developed solutions by other teams, followed by the applicable
rules imposed by Formula Student competitions to wheel design.
2
Parts four, five and six will constitute the core of the document. In part four we will
analyse the dynamic behaviour of a race car and the effect a reduction in non-sprung,
rotating mass will have on its on-track performance. This will provide the basis for a
Cost vs. Performance analysis using the Key Efficiency Metric, a tool developed by MAD
Formula Team for this purpose, which will then be used to set a budget cap for the project.
In part five, the forces and reactions acting upon the rim will be analysed in order to
develop a base FEA model that will later be used to simulate and optimize the design
of the rim, including the analysis of the mechanical properties of the materials used,
specifically the short-strand, compression-molded carbon fiber. Part six will cover the
design of the rim itself, as well as the simulation through FEA software, and its adaptation
for low-budget manufacturing, including whether or not the manufacturing of the rims
can be achieved within the set budget, with a conclusion to the work developed in this
project. Parts five and six were developed simultaneously, but have been separated to
make reasoning, calculations and decisions more clear to the reader.
Finally, part seven will offer the conclusions to the project, while part seven will in-
clude a roadmap of future developments, including the adaptation of the low-cost manu-
facturing process for series manufacturing, and the democratization of the access to CFRP
rims by low-budget teams via parametrization of key design variables on the rim.
Additional documentation will be provided in the form of an annex, which will include
material datasheets and other relevant information.
1.4. Tools and Software used.
The following software and software suits were used for the development of this work:
• Finite Element Analysis Pre-processing: Altair Hypermesh.
• Finite Element Analysis Solver: Altair Optistruct.
• Finite Element Analysis Post-processing: Altair Hyperview.
• Computer Aided Design: Dassault Systemes SolidWorks.
• Product Lifecycle Management/CAD repository: StrataSYS GrabCAD.
• Mechanical properties analysis, data analysis, graph-making, budget control: Mi-

crosoft Excel.
• Laptime Simulation: OptimumG OptimumLap.
• Project management, Key Efficiency Metric: MADFT Project Master Document.
3
2. SOCIOECONOMIC ENVIRONMENT: FORMULA STUDENT.
2.1. Brief history of Formula Student competitions.
Formula Student is a university-level engineering competition in which teams of undergraduate-

and masters-level students, mostly from engineering disciplines, design, build and race a
single-seater, open-wheeled (commonly known as "Formula-style") racecar.
The first competitions where organized by the Society of Automotive Engineers in
the USA back in 1980, and quickly gained traction within North American technical
universities, with over 35 teams registered by 1985. Known as Formula SAE or FSAE,
and hosted by different universities, they attracted the attention of industry giants such as
Ford and General Motors, and also professional racing teams at all levels.
Fig. 2.1. University of Texas-Arlington 1986 champion FSAE race car, courtesy of the University
of Texas-Arlington News Service photograph collection.
In 1998, FSAE made it to Europe with the first Formula SAE UK competition, attract-
ing interest from the top European engineering schools, and reaching worldwide popular-
ity in 2006 with Formula SAE Australasia, held in Australia, which finally allowed teams
from all over the world to compete with their local peers.
4
Fig. 2.2. University of Birmingham 1998 race car on the first Formula Student UK competition,
courtesy of Racecar Engineering Magazine.
In 2005, the first Formula Student Germany competition was held, setting up the table
for what is commonly known as the European Season, which nowadays comprises ten
competitions in ten different nations and spans from early July all the way into September,
resulting in almost three months of uninterrupted, back to back competition.
To promote new research, development and design, as well as levelling the play-
ing field and improving safety and competitiveness, the rules for the competitions have
changed on a yearly basis. The biggest changes occurred in 2010, with the introduction
of the Electric Vehicle (EV) class, 2016, with the introduction of the Driverless Vehicle
(DV) class (albeit as a concept-based class, with the first true driverless cars not racing
until the following year), and 2017, with the introduction of the Formula Student Ger-
many rules, which replaced the FSAE rules for the competition. The FSG ruleset was
quickly adopted by the European competitions, with the exception of Formula Student
UK and Formula SAE Italy, which allow cars built under both sets of rules to compete.
The FSG ruleset is regarded as more advanced and innovative, including provisions for
more advanced aerodynamic devices, fully- or partially-autonomous vehicles, and higher
safety standards.
5
2.2. The 2022 Formula Student Germany rules and its derivative rulesets.
Formula Student Germany is the most prestigious international Formula Student compe-
tition, as well as the organization responsible for developing the ruleset for all European
competitions, except FS UK and FS Italy, which operate under FSAE rules, but allow cars
legal under FSG rules to participate, too.
Formula Student Germany began developing its own set of rules in 2017, once it
became clear that the FSAE rules, developed by the American Society of Automotive En-
gineers, could not accommodate the realities of the European competitions, which tended
to showcase more advanced prototypes, including, for the first time, fully functional au-
tonomous racecars (under the Driverless Vehicle category, which began as a separate class
that same year).
The FSG rules are mostly focused on the safety of all participants, and set the mini-
mum baselines for a number of systems, such as the chassis or safety electronics. They
also establish an upper ceiling to performance, aiming to keep the amateur drivers out of
harms way and to level the playing field by imposing limits to engine power (both electric
and from internal combustion engines) and aerodynamic devices.
After the 2019 competitions, Formula Student Germany announced a new roadmap,
interrupted by the COVID-19 pandemic, including the gradual phase-out of the Combus-
tion Vehicle Class and the introduction of the new Driverless Cup, making a requisite for
participation that all cars have at least partial self-driving capabilities by 2022, delayed to
2023 to account for the cancelled 2020 season. The move was widely unpopular among
the European and international Formula Student team, who felt it was motivated by ex-
ternal pressure from sponsors and the automotive industry, and feared that it would push
smaller teams out of the competition, as development of both Electric and Driverless Class
vehicles is more expensive by an order of magnitude. In response, other European compe-
titions, such as FS Austria, FS Czech and FS East (Hungary) released a new set of rules,
derived from the FSG rules, but eliminating the Driverless requirements, and for the first
time allowing Teams to integrate hybrid systems on their internal combustion-powered
vehicles.
2.3. Formula Student Events and Scoring.
Formula Student competitors are scored both on and off the track, out of a total possible
score of 1000 points [“Rules 2022”, 2021]. On Static Events, worth a total 325 points, a
panel of expert judges will analyse either the design, manufacturing or business planning
capabilities of the team; whilst on Dynamic Events, worth the remaining 675 points, the
teams will be scored by the performance of their car on track on four separate disciplines.
In order to attend a competition, teams must participate in qualifying quizzes, usually
held over the course of one day on the last week of January. Quizzes include questions
6
on basic engineering knowledge, specifics of race car engineering, knowledge of the For-
mula Student rules, and oftentimes a random mix of questions relating to local culture
of the host nation, pop culture, inside jokes relating to the Formula Student community
and bits and pieces of trivia.Teams are ranked by the number of correct answers and the
time needed to complete the quiz, with only the top teams being eligible for a slot at the
competition.
All Formula Student competitions begin with a Technical Inspection, which will de-
termine whether the prototypes are safe and compliant with the rules. Inspection order
is slot-based, with slots typically assigned randomly. The inspectors will assess each car
following a strict checklist, divided into five segments for Combustion Class vehicles:
• Pre-scrutineering: Checks to the driver safety gear (restraints, fireproof underwear

and suits, helmets, gloves and boots), as well as other required items such as dry-
and wet-weather tyres, fire extinguishers and specific tools.
• Scrutineering (Mechanical): Fulfillment of general rules, safety and structural in-

tegrity of the car, including chassis, suspension, aerodynamics, and powertrain.
• Tilt test: Rollover stability of the car with the tallest driver at 30º and 60º angles,
plus spillover of fluids from the cooling, lubrication or fuel storage systems. The
car is then weighted.
• Noise test: Compliance with local noise restrictions, as well as checks to the elec-
tronic safety system and intake diameter restriction.
• Brake test: On-track, the car must demonstrate capability to brake locking all four
wheels simultaneously and without stalling out.
After each segment is completed, an Inspection Sticker is applied to the car. These stick-
ers serve as a visual indication that the car has demonstrated compliance with the required
inspections, and can be removed at the discretion of competition personnel if for any rea-
son the car becomes, or is suspected to be, non-compliant. Passing the relevant Inspection
again will be required to recover the sticker and progress through the competition again.
Once the prototype has passed the Mechanical Inspection, it is allowed to access the
practice area and can take part on the dynamic disciplines, which are:
• Manual Acceleration Event (50 points): From a standing start, the vehicle must
accelerate over 75 meters. Four runs, driven by two different drivers, are allowed.
The aim of the event is to test the ability of the car to accelerate on a straight line
and maintain traction and control. 3.5 points are scored by all teams who complete
a successful run, and additional points are scored following the formula on figure
2.3.
7
Fig. 2.3. Scoring for the Manual Acceleration Event, FSG rules 2022.
• Manual Skidpad Event (75 points): From a rolling start, the vehicle must complete
a figure-eight circuit as shown on figure 2.4, with two laps on the right-hand cir-
cle followed by two laps on the left-hand cycle. Each second lap is timed and the
average of the two timed laps is the result for the event. Four runs, driven by two
different drivers, are allowed. The aim of this event is to test the steady-state cor-
nering ability of the racecar. As in the Acceleration Event, 3.5 points are scored by
all teams who complete a successful run, and additional points are scored following
the formula on figure 2.5.
Fig. 2.4. Layout for the Skidpad figure-eight circuit, FSG rules 2022.
Fig. 2.5. Scoring for the Manual Skidpad Event, FSG rules 2022.
• Autocross Event (100 points): In the Autocross Event, the cars must complete a lap
of a handling circuit, with an length around 1 km, in the shortest possible time. The
circuit may include elements such as constant- and variable-radius corners, straights
under 80 meters in length, hairpins (180º turns) with a minimum radius of 9 meters,
and slalom sections, plus variations in track surface and changes in elevation due to
8
the terrain conditions. Four runs can be attempted, driven by two different drivers,
and every team capable of achieving at least one successful run will be rewarded
with 4.5 points, with additional points scored following the formula on figure 2.6.
Furthermore, the best time for the Autocross Event will serve as the qualifying time
for the running order on the Endurance Event.
Fig. 2.6. Scoring for the Autocross Event, FSG rules 2022.
• Endurance Event (325 points): The longest and most difficult Event on a compe-
tition, the Endurance tests the ability of the Team to develop a fast and reliable
racecar. Over the course of 22km, the prototypes will lap around a track based on
the Autocross circuit but slightly modified to allow for passing lanes, as there may
be several cars running at the same time. Teams will be ranked according to their
fastest Autocross times, so that only cars with roughly the same capabilities will
be lapping the circuit at the same time, minimizing risks (although cars can be al-
lowed to run out of order, with a time penalty). There is only one run allowed for
this Event, with a mandatory pit-stop and driver change at the mid-distance mark.
Teams that manage to successfully complete their run will be rewarded with 25
points, plus additional points following the formula on figure 2.7.
Fig. 2.7. Scoring for the Endurance Event, FSG rules 2022.
• Fuel Efficiency (75 points): Fuel Efficiency is part of the Endurance Event, but
scored separately by all teams who complete the Endurance Event with a maximum
fuel consumption of 15kg/100km for cars running on RON98 fuel or 21.75kg/100km
for cars running on E85 fuel, and manage to complete their endurance run within
1.333 times the final endurance time for the fastest vehicle. The goal of the Event
is to demonstrate the capability of the Teams to develop a racecar which is fuel-
efficient but can still achieve competitive lap times. There are no points scored for
a successful run, with the entire scoring of the event given by the formula on figure
2.8.
9
Fig. 2.8. Scoring for the Efficiency Event, FSG rules 2022.
For the 2022 season, mandatory Driverless Skidpad and Acceleration were required to
opt for the full 1000 points, as the typical score of 75 points for Skidpad and Acceleration
was capped at 50 points for the manual Events.
The Static Events typically occupy the first few days of the competition, and are often
mixed in-between the Technical Inspection slots, helping free up the Technical Inspection
and smooth out the running of the competition, and can score points for the teams even
if the Technical Inspections are unsuccessful. The teams will defend their choices and
developments and receive feedback from industry experts and professionals from disci-
plines related to automotive design, business management and manufacturing. The Static
Events are:
• Engineering Design (150 points): Teams will have to defend their design decisions,
backed up with data from simulations, models and real-world testing, in order to
justify the design and development of their current prototype. Factors such as team
management, transfer of knowledge and resource allocation will be considered, as
well as understanding of basic engineering principles, concept application and good
engineering practices. Teams will be ranked and scored as per figure 2.9, with the
top teams progressing to a final judging round.
Fig. 2.9. Scoring for the Engineering Design Event, FSG rules 2022.
10
• Cost and Manufacturing Event (100 points): Teams will have to show understand-
ing of the manufacturing process for their prototype, including project management,
cost estimation, and environmental aspects, as well as being able to discuss how dif-
ferent manufacturing methods for both series and prototype production would affect
the expected performance of the car. Teams will be ranked and scored as per figure
2.10, with the top teams progressing to a final judging round.
Fig. 2.10. Scoring for the Cost and Manufacturing Event, FSG rules 2022.
• Business Plan Presentation (75 points): Teams will have to present a viable business
plan, based around their prototype, to a panel of judges who will act as potential
investors. The teams will have to show a good understanding of fiscal and financial
concepts, as well as long term financial planning and business acumen, in order to
convince the investors to back their idea. Teams will be ranked and scored as per
figure 2.11, with the top teams progressing to a final judging round.
Fig. 2.11. Scoring for the Business Plan Presentation Event, FSG rules 2022.
2.4. MAD Formula Team
MAD Formula Team is the Formula Student team from Carlos III de Madrid University.
Founded in 2011 as Escudería UC3M, the team took part on competitions under the name
Formula UC3M, often stylized as FUC3M, before rebranding as MAD Formula Team.
As Formula UC3M, the team took part in four events (2016, 2017, 2018 and 2019) at
Formula Student Spain and one event (2019) at Formula Student Germany, receiving no
accolades. MAD Formula Team has taken part in two events (2021 and 2022) at Formula
Student Spain, two events at Formula Student Austria (2021 and 2022), one event at
Formula Student Germany (2022) and one event at Formula Student Netherlands (2021),
and obtainede two Combustion Vehicle Class victories for the Cost and Manufacturing
11
Event, in FS Netherlands 2021 and FS Spain 2022, as well as a second place for the
Skidpad Event (FS Spain 2022), and a third place for the Engineering Design Event in FS
Spain 2022.
For the first car ever developed, the C3R16 from 2016, the team introduced a full
aerodynamic package on a tube frame chassis, which paired with a four-cylinder Honda
engine made for an extremely heavy vehicle. Despite the weight of the car exceeding
300 kilograms, outright doubling the weight of the top contenders from other universities,
the C3R16 prototype was able to clear Technical Inspection (no mean feat for a first
prototype by any metric) and run all dynamic events, although it was unable to complete
the Endurance event. In any case, it provided a much needed first platform to improve on,
as well as a morale booster for the budding team.
Fig. 2.12. Formula UC3M C3R16 in Montmeló, Catalonia, Formula Student Spain 2016.
Courtesy of MAD Formula Team.
For the following prototype, a sponsorship from FIDAMC (Fundación para la Investi-
gación, Desarrollo y Aplicación de Materiales Compuestos) allowed the team to develop
the first hybrid chassis, with a composite monocoque for the front half of the car, includ-
ing the cockpit, and a tubular subframe for the engine bay and rear suspension, in an
effort to reduce the weight of the vehicle whilst retaining the full aerodynamic package
and four-cylinder engine. Once again, the car cleared Technical Inspection and was able
to run, but not finish, all dynamic events.
12
The third car, C3R18, was a further improvement on the 2017 platform, with a new and
improved hybrid chassis that was lighter and stiffer than before. The car run flawlessly for
the entire competition, passing Technical Inspection and competing in all dynamic events,
but an engine failure in the final laps of the Endurance event sadly ended its chances at an
excellent result.
In 2019, the team made took part in its first competition abroad at the prestigious For-
mula Student Germany event, after an excellent result in the registration quizzes, as well
as Formula Student Spain. However, a slew of manufacturing issues plagued the devel-
opment of the car, resulting in not passing the Technical Inspection at Formula Student
13
Germany, and thus not running in any dynamic events, and barely passing the Technical
Inspection at Formula Student Spain, being able to barely run a full lap of the Endurance
event before failure.
Fig. 2.15. Formula UC3M C3R19 in Hockenheim, Baden-Württemberg, Formula Student

Germany 2019. Courtesy of MAD Formula Team.
Following the 2020 COVID-19 pandemic, Formula UC3M was officially dissolved
and refounded as MAD Formula Team, following a rebranding and corporate image
changes, as well as a change in Team philosophy, opting for a simpler, more well-rounded
design approach as well as a better understanding of the underlying engineering principles
behind racecar design and manufacturing, always looking to find and understand the lim-
its of the regulations. Despite the difficulties of developing a racecar with what effectively
was a completely new team during a global pandemic, the Team obtained their best results
to date on the competition qualifying events and registered for three international com-
petitions, Formula Student Netherlands, Formula Student Austria and Formula Student
Spain. With a car completed in near-record time and the new organizational structure,
MAD Formula Team was able to obtain their best-ever results with a combustion-class
victory on the Cost and Manufacturing event at Formula Student Netherlands, as well as
top 10 and top 5 finishes in several static and dynamic events in Austria and Spain.
14
Fig. 2.16. MAD Formula Team MFT01 at the car rollout event, Leganés, Madrid, 2021. Courtesy
of MAD Formula Team.
On the recently finished 2022 season, MAD Formula Team achieved the largest suc-
cess in its short history, first by classifying for all European competitions (being the only
team in Europe to do so) and then by obtaining three podiums, two in Static disciplines
(Cost and Manufacturing, and Engineering Design) as well as the first Dynamic Event
podium for the Team, with a second-place finish on the Formula Student Spain 2022
Skidpad Event.
15
3. STATE OF THE ART.
3.1. Commercially-developed rims.
Commercially available CFRP rims can be separated into three broad categories: rims
for commercially available cars, rims for motorcycles, and rims designed specifically for
Formula Student and Formula SAE vehicles, available from select manufacturers.
While full CFRP wheel rims have not yet become normal on road going vehicles,
where the state of the art is lightweight alloy rims, some manufacturers do offer the op-
tion to specify them for select models of high-performance, track-oriented cars. Such
options have high price tags attached, as they are both a performance upgrade and an
exclusive product, driving the prices up. At the time of writing, the only manufacturers
who make CFRP wheel rims available as a factory option would be Porsche AG (with a
price of $14980), Ferrari ($15000) and boutique supercar manufacturer Koenigsegg (at an
estimated cost of between $40000 and $50000). CFRP rims can also be purchased from
aftermarket brands, such as Carbon Revolution, who manufacture them for a variety of
car makes and models, such as Ford, Audi or Renault. Whilst prices vary, they are typi-
cally above the $10000 mark, and in most cases the manufacturer will not offer a quote
until they know the exact number of rims ordered.
In terms of technology, most manufacturers have kept their process secret, in order to
retain a competitive edge on the market. Koenigsegg, however, published a short video
[“Making 280mph capable carbon fiber wheels”, 2014]in 2014 showcasing part of the pro-
cess, which depicts a technician manufacturing one of their CFRP rims while the owner
of the company talks the viewer through the steps required. On the video, the rim is
manufactured using pre-impregnated woven carbon fiber cloth, laid into a negative alu-
minium mold and cured. Whilst the video is informative, the high price of the machined
aluminium molds makes low-cost manufacturing, one of the key points of this project,
unachievable.
16
Fig. 3.1. Koegnisegg rim manufacturing, from the THE DRIVE youtube channel.
CFRP rims for motorcycles are somewhat more common, derived from the need to
reduce rotational inertia from the wheels to decrease steering effort, as motorcycles need
to lean into a turn, a movement opposed to by the gyroscopic effect of a rotating wheel and
dependent on the mass of the rotating assembly. However, their manufacturing process
is also unsuitable for our project, as motorcycle wheels are woven from carbon thread,
braided using specific machinery and tooling before curing. The only notable mention
would be Rotobox [ROTOBOX, 2022], using what they call Forged Carbon, a marketing
term for CMCFRP. In terms of price, motorcycle CFRP wheel rims are cheaper than
their car counterparts, mostly due to their more widespread availability, driving a more
competitive market, and to the lower mechanical demands of motorcycles.
Finally, there is one model of commercially available CFRP wheel rims specifically
for Formula Student prototypes, made by BRAID. The BRAID Sturace Carbon Fibre
[BRAID, 2022], available in just one size (7x10", albeit with different offsets and mount-
ing configurations), with a cost per rim of between $2254 and $2401 depending on the
configuration. With a price per set of four of, at a minimum, $9016, and at least two set
required as per the Formula Student rules (demanding one set of dry-weather tyres and
one set of wet-weather tyres mounted on rims and always ready to use), commercially
available Formula Student CFRP rims cannot be considered as a low-cost option, either.
17
Fig. 3.2. BRAID StuRace carbon, from the BRAID Website.
3.2. Student-developed rims as state of the art.
Considering the high prices of commercially available products, and the specific dynamics
of a Formula Student race car, covered on section 4 of this document, many teams have
tried, and succeeded, in developing their own CFRP wheel rims.
Self-developing rims allows the teams to fix their own targets in terms of weight and
cost, with the lightest example of a CFRP rim was by TU Graz (University of Graz,
Austria) in the 2019 season, at roughly 600 grams per rim. The team achieved that result
by developing a two-part rim, which allowed them to hollow the spokes and further reduce
the weight of the assembly. Both halves of the rim were laminated separately and then
bonded with adhesive and riveted together.
Fig. 3.3. TU Graz rims and cross section, courtesy of TU GRAZ Racing Team.
Although the resulting rim was quite impressive, it suffered a catastrophic failure dur-
18
ing the Endurance Event at Formula Student Spain 2019, resulting in a DNF. Since then,
the TU Graz FS team has moved over to electric cars with in-hub gearboxes and thus
adapted the design for their new requirements, and adopted a similar solution to TU Delft
(University of Delft, Netherlands), which will be covered a bit further into this section.
UAS Graz (JOANNEUM University of Graz, Austria) adopted a similar solution, al-
though they did not achieve such an impressive result in terms of weight, since their
wheels were not hollow. Instead, they laminated the rim in one piece, using CNC ma-
chined foam core for their spokes.
Fig. 3.4. UAS Graz rims, courtesy of FSG Media Team.
Older designs by FS teams such as KIT Racing (Karlsruhe Institute of Technology,

Germany) or ZIPS Racing (University of Akron, USA) [Kruse et al., 2021] show a con-
vergence towards the three spoke rim design, although in their case the full spoke is made
of monolithic carbon. The design can be assumed to be extremely reliable, as such a large
monolithic element can withstand enormous forces, but offers reduced benefits in terms
of weight compared to a metallic rim.
19
Fig. 3.5. University of Akron rims.
JMS Racing (University of Kansas, USA) [Walther, 2016]developed a simple rim

with five hollow spokes. The design requires aluminum molds and rubber tooling, which
is extracted from the half-finished rim before final lamination.
Fig. 3.6. University of Kansas rims.
Despite initial results on a test bench being successful, JMS Racing decided against
using this design after track testing, opting for a design more in line with TU Delft Bucket
and Crown design. In this design, the main part of the rim is a CFRP bucket, that is, a
simple cylinder open on one end, which is attached to the rest of the wheel and suspension
assembly via a metallic crown which acts as the spokes on the rim. This design allows
for a greater freedom when designing the upright, which is why it has become a relatively
common design for EV with inboard gearboxes. The most extreme case is, of course,
TU Delft, which uses a sintered titanium upright and crown to mount their gearbox, and
20
mounts their brakes on the outboard side of the wheel for improved cooling efficiency.
Fig. 3.7. TU Delft rims, courtesy of FSG Media Team.
3.3. Why student-developed rims are the state of the art in FS.
Student-developed rims doubtlessly constitute the state of the art in Formula Student.
With no pressure to make sales and no regulatory constraints, Formula Student teams have
been able to push past any widespread rim technology to obtain rims that are extremely
lightweight and perfectly adapted to their suspension and wheel assembly packaging.
Even FS-specific commercial rims, such as the BRAID Sturace, have to be designed for
oversized loads, lest the rims fail during normal operation and the affected teams make
a claim against the company causing damage to their finances and reputation, resulting
in rims effectively lighter than their aluminium or magnesium counterparts, but still very
heavy compared to some of the lightest team-developed rims.
Self-developed rims also entail a significant reduction in costs for the team. Even
though the initial cost is high, as both the material and machining for molds is very ex-
pensive, once the team has the manufacturing capability to make their own rims there
are no limitations to the number of rims that can be produced. The cost of new rims is
reduced to the cost of materials, facilities and labor, which in terms of Formula Student
can be practically null, as most teams have sponsors covering the first two and use team
members as volunteer labor.
Furthermore, new manufacturing methods, including rapid prototyping, 3D printed
molds and CMCFRP can further reduce the cost of materials and tooling and even make
use of recycled materials possible.
21
3.4. Regulatory Framework: Formula Student regulations pertaining wheel rims.
The Formula Student rulebook [“Rules 2022”, 2021] does not provide any regulations on
rims or rim design, focusing entirely on the mechanical fasteners attaching the rim to the
hub assembly, with every paragraph of rules article T 2.5 as follows:
• Article T 2.5.1, regarding wheel rims mounted with single nut (also known as Cen-
terlock) and their retaining systems.
• Article T 2.5.2, regarding wheel rims mounted on lugs or studs, restricting the ma-
terials to steel and titanium, construction to solid lugs, and requiring calculations to
demonstrate adequate strength.
• Article T 2.5.3, regarding the use of aluminium retaining nuts in pristine, hard-
anodized condition.
Fig. 3.8. Article T 2.5 of the 2022 Formula Student Germany Rules.
There is one more rule affecting wheel rims, pertaining to the Technical Pre-inspection
(Pre-scrutineering), as mentioned on section IN 2.1, where technical inspectors will
check that the team has all required equipment to run their car safely. During this pre-
scrutineering, inspectors will check that the team has two sets of tyres mounted on rims,
and will inspect the rims for damage or manufacturing issues (article IN 2.1.1).
Fig. 3.9. Article IN 2.1 of the 2022 Formula Student Germany rules.
22
In terms of materials, shape and size for the actual rim, the FSG regulations give al-
most total freedom to the designers, provided that they follow good engineering practices
and can show appropriate mechanical properties to withstand the loads on the Dynamic
Events. With that freedom, the design of the rims can be carried out following the re-
quirements set by the Team. Concerns over international rim standards that apply to the
automotive industry can be disregarded, as the wheel rims will not be available for road
use, will only be used under strict safety conditions on closed tracks, and do not require
any certification or approval process.
23
4. DYNAMIC ANALYSIS OF THE IMPACT OF CFRP RIMS ON
RACE CAR PERFORMANCE.
4.1. Basics of FS Dynamic disciplines.
As discussed on section 2.2, Formula Student dynamic disciplines constitute 67.5% of

all available points of a Formula Student competition, and are therefore of the utmost
importance.
In Formula Student, with the exception of the Acceleration Event, circuits tend to be
small, narrow and twisting, with many turns and very few straightaways, and top speeds
of around 85km/h. Cornering is, then, where the car is expected to spend most of its time
and therefore the most critical phase in any dynamic event. In order to properly asses
the importance of weight reduction for unsprung, rotating mass, we have to define the
cornering process and separate it into three segments, which will already be familiar to
most people with an interest in automotive racing: braking, turning and accelerating.
Assuming an arbitrary corner of any shape and radius, at the end of a straightaway
of any length, the car will approach at a speed which can be safely assumed to be higher
than the maximum speed at which the car can take that corner, defined by the mechanical
and aerodynamic grip (which we will, for now, assume to be a fixed and known amount,
and discuss further on section 4.2). In order to make the corner, the driver must brake
in a straight line, before turning the car into the apex maintaining the maximum possible
momentum. Finally, after the apex, the driver will begin to turn the car out of the corner,
carefully accelerating so as to not lose control of the vehicle. Seasoned drivers will trail-
brake, that is, gradually trail-off the brakes as they turn into the apex and gradually trail-
on the throttle as they turn out of the apex, following an optimal path through the corner,
the racing line. The narrow nature of Formula Student circuits and the amateur nature of
Formula Student driver makes the optimal racing line a non-issue, but some fundamentals
of auto racing still apply: drivers must brake as late as possible, maximizing the time
running at full-on throttle; carry as much velocity as possible through the turn, where they
have to carefully balance the car at the edge of grip; and accelerate as soon as possible
without losing control or spinning out [Seward, 2014].
The ability of a car to brake later into a corner, maintain more speed through the
corner and accelerate out of the corner sooner than its rivals makes the difference between
a leading car and a backmarker, even if the rival cars have more engine power, higher top
speed or more efficient aerodynamic devices.
24
Fig. 4.1. Formula Student Germany race track, Formula Student Germany handbook. Notice the
abundance of tight corners and lack of straight sections. 25
4.2. Basic dynamics of a race car.
Lower weights for all elements of the car can help achieve the target results: assuming
that the forces, delivered by the brakes and powertrain, acting upon the car remain equal,
a lower mass directly translates into higher longitudinal acceleration, both positive and
negative, corresponding to the acceleration and braking phases of cornering. This can be
trivially shown from Newton’s Second Law.
Mass, sprung or unsprung, is not a deciding factor in the cornering ability of the car,
which is governed by the balance between the maximum grip of the car, a result of the
mechanical and aerodynamic grip of the tyre, pulling the car into the corner, and the
centrifugal force, pushing the car out of the corner. Although reducing the mass of the
car does not directly affect the turning capabilities of the car, the reduction of unsprung
mass makes the suspension work more efficiently, maximizing the contact between tyre
and surface and improving grip, which directly translates into better traction during all
cornering phases and higher top speed at the apex.
Reduction to the rotational inertia of the car is also one of the key reasons to devote
the time and resources to develop lighter rims. High rotational inertia reduces the braking
and acceleration capabilities of the car as it opposes any changes to the angular velocity
of the wheel, and also affects the turning ability of the car, as the gyroscopic effect of
the rotating mass tries to keep the wheel aligned with its axis of rotation, an effect that
is exacerbated by the comparatively high camber on race cars, as opposed to normal,
road going cars. The rotational inertia of the rims also affects the efficiency of the entire
powertrain, as the rims are one more element on the cinematic chain from the engine,
through the transmission, differential and driveshafts and unto the tyre and road. [Seward,
2014]
4.3. Lap Time Simulation. Introduction to the Key Efficiency Metric.
Lap Time Simulation, or LTS, is a tool used in automotive racing design to make deci-
sions regarding design and set up. Since real world testing is costly in terms of facilities,
personnel and wear and tear on the car and its components, and many series restrict the
testing available to competitors, developing an accurate computer model of the car to
simulate different options can help the team make decisions.
High-echelon motorsport series, such as Formula 1 or World Endurance Champi-
onship, make extensive use of lap time simulation tools, developed by the different teams
and with their full capabilities kept a closely guarded secret. They typically set up their
code to run thousands of full-race length simulations, sweeping through a broad range of
parameters such as fuel load, tyre strategy, aerodynamic and suspension set up changes,
engine mapping, or energy use, and including outside events such as race incidents, in-
terruptions, or driver fatigue. Obviously, FS lap time simulation is not nearly as capable.
26
Most teams have to either develop their own lap time simulation code or rely on com-
mercial software alternatives. For this project, the LTS software used was OptimumLap,
kindly provided free of charge by software developer OptimumG.
OptimumLap offers, as part of its packaged, a predefined FSAE-style car, with or
without the aerodynamic elements. In both cases, a set of parameters, including weight,
engine power and transmission ratios, powertrain efficiency and fuel type, thermal effi-
ciency, and tyre data, are provided, and for the version with aerodynamic elements, lift
and drag coefficients. Although the expectation is that a team will set those parameters
to match their actual racecar, for the purposes of facilitating the analysis of the perfor-
mance impact of CFRP rims we can simply take the provided values for all variables ex-
cept weight and powertrain efficiency, and use a two-parameter batch sweep between the
minimum and maximum values expected from a change between MAD Formula Team’s
current set up (cast aluminium alloy rims) and all possible configurations of CFRP rims.
With a weight of 2.5 kg for cast alloy rims and anywhere between 600 g to 1.3 kg ob-
served on CFRP rims, the weight of the car in OptimumLap will be set to sweep between
the default value (230 kg) and a minimum value of 222.4 kg, corresponding to the weight
of the car with the lightest possible rim configuration.
A second sweep parameter, drivetrain efficiency, will account for the reduction on
rotational mass, from the default 80% to a maximum of 85%, corresponding to a reduction
of the rotational mass of the drivetrain of around 15%.
Both sweep parameters were incremented first over 15 steps and then over 20 steps,
the maximum allowed by the software, in order to verify the linearity of the results.
Results are shown on figures 4.2, 4.3, 4.4, 4.5, 4.6 and 4.7 for one lap of a typical
Autocross/Endurance Event circuit, such as the one on figure 4.1 (in this case, Formula
Student Germany Autocross/Endurance Event circuit, kindly provided by OptimumG),
for three parameters: total lap time, average speed, and fuel use.
27
Fig. 4.2. Laptime vs. Weight, Powertrain efficiency sweep chart, 15 steps. Lower is better.
Fig. 4.3. Laptime vs. Weight, Powertrain efficiency sweep chart, 20 steps. Lower is better.
28
Fig. 4.4. Average speed vs. Weight, Powertrain efficiency sweep chart, 15 steps. Higher is better.
Fig. 4.5. Average speed vs. Weight, Powertrain efficiency sweep chart, 20 steps. Higher is better.
29
Fig. 4.6. Fuel consumption vs. Weight, Powertrain efficiency sweep chart, 15 steps. Lower is
better.
Fig. 4.7. Fuel consumption vs. Weight, Powertrain efficiency sweep chart, 20 steps. Lower is
better.
From the LTS data it can be clearly discerned that most of the benefits come from
the reduction in weight for the car, with only a small improvement correlating directly
to the increased efficiency of the drivetrain, except for fuel consumption. Nonetheless,
considering that fuel efficiency can account for up to 7.5% of the total available points in
a competition, the effects cannot be disregarded. The differences between the 15- and 20-
step sweep, however are effectively negligible.
Once the performance gains have been accounted for through simulations, the results
are introduced into MAD Formula Team’s self-developed project management tool, the
Project Master Document or PMD, an Excel-based system integrating all management
30
functionality for the team. As part of the software package, the Team has developed the
Key Efficiency Metric or KEM, a decision-making tool which will set a target budget for
the project to be considered viable.
The KEM can take as inputs the expected points differential between two different
options for every scoring part of the competition, the required budget for each of the
options in terms of both money and manpower, and coefficients depending on how likely
a project is to succeed beyond the Team’s best expectations or fail below the Team’s worst
estimates. Although the tool is meant to provide an objective, reliable metric, some of the
data from inputs has to be estimated using team members’ experience, while other can be
analyzed from previous competition data. In the analysis for section 4.4, only objective
competition data and results from simulations will be used.
4.4. Project Budget: Impact and viability analysis.
Taking the results from the Formula Student Germany 2022 dynamic events, the points
delta between three different configurations can be calculated: the best time by MAD
Formula Team, using the current 2.5 kg aluminium alloy rims; a mid-point time, corre-
sponding to CFRP with a weight around 1.25 kg, half of the weight of the current alloy
wheels; and a best time, using the lightest possible rims at 600 g. For comparison, the
best time by MAD Formula Team for the 2022 Autocross Event was of 97.630 s, with a
car weighting 232 kg, very close to the parameters of the sweep, and the best time, by
UAS Esslingen (Esslingen University of Applied Sciences, Germany) was 77.744 s.
LAP TIME [s] SCORE [pts] POINTS DELTA TO BEST [pts]

UAS Esslingen 77.744 100 0
MADFT, Alloy Wheels 97.630 4.5 95.5
MADFT, 600g CFRP Wheels 84.750 60.52 39.48
MADFT, 1250g CFRP Wheels 84.810 60.21 39.79
Table 4.1. BEST TIMES, CORRESPONDING SCORE AND POINTS

DELTA TO THE BEST SCORE AT FSG2022 AUTOCROSS EVENT
FOR DIFFERENT WHEEL RIM CONFIGURATIONS.
The results on Table 4.1 clearly show that although there is a very significant dif-
ference between the current alloy rims and any CFRP rims, once the weight of the rim
assembly is below 1250 grams, the differences in points can be negligible. Considering
that the cost of manufacturing rims grows as the weight decreases, and the delicate fi-
nancial situation of MAD Formula Team, focus on development of the lightest possible
rims would be inadvisable: development should be oriented towards cost, and not weight,
optimization.
In order to perform the KEM analysis, points delta for Dynamic events were consid-
ered to be in line with the results obtained by the simulations, that is, considering that the
31
Team will achieve roughly 60% of all available points. Although this procedure is un-
orthodox, it reduced the time required for the preliminary analysis of viability for CFRP
rims, leaving more time for the design and analysis of the actual rims.
Points from the Endurance and Fuel Efficiency Events were not accounted for, as the
2022 MADFT race car was, unfortunately, unable to complete the Events, and therefore
any scoring would taint the results of the analysis, whilst points delta for Static Events
were also disregarded, as the estimation for this scoring relies on team member’s expe-
rience and could be tainted by the author’s personal bias and opinions. In any case, this
was a more restrictive condition, as the lap time simulations were showing improvements
on both lap times and fuel efficiency, which would have increased the Endurance and
Fuel Efficiency scores; and showing a new design of CFRP rims, manufactured with a
rather uncommon process, would likely increase the scoring for the Design and Cost and
Manufacturing Events.
The KEM, shown in figure 4.8, takes the expected points delta between two config-
urations, as well as the expected budget for each option and the required manpower, as
a function of number of people working on a project, the required hours per person per
week, and the number of weeks required for development. Additionally, two compensat-
ing factors can be added, an optimist (compensating for how likely it is that a development
may cause a larger points increase than expected) and a pessimist (compensating for how
likely it is that a development may cause a smaller points increase than expected) factor.
Since both factors are subjective, they were not used for this project.
Fig. 4.8. Empty KEM tool, before data entry.
The KEM takes the above data as inputs and outputs two metrics, known as KEMA
and KEMB , one for each option. The metric has two parts, an absolute number and a
percentage. The absolute number represents the points gained, or lost, by the develop-
32
ment adjusted for the money and work required to implement, whilst the percentage is a
decision-making tool representing how sure the software is that the decision to implement
a certain development is correct.
Working backwards, the KEM can define a maximum budget for money and work
hours to be dedicated to a development before it is no longer justifiable to implement.
That is achieved by inputting the expected points delta and adding or subtracting money
and work hours until the equilibrium point of 0.0 and 0.00% for each part of the metric is
achieved. Since the hours to dedicate for this project are fixed by the requirement that the
work be developed in 12 ECTS time units, corresponding to 300 hours, over 16 weeks,
the only variable is money. On figure 4.9, the equilibrium point has been reached, with
the final budget allowed for the project being €4281.
Fig. 4.9. Balance point on the KEM metric and final budget.
Considering that some of the costs will be sponsored as part of the MADFT yearly
development of a racecar, two budgets are provided on figure 4.10: one where the cost of
sponsored materials and facilities is not included, and one that accounts for non-sponsored
cost of the entire process.
The real cost of developing and manufacturing CFRP rims is so beyond the reach of
MAD Formula Team that it is a laughable endeavour. However, sponsored manufacturing
can be viable using a two-mold fabrication, with 3D printed molds, within the budget
allowed by the KEM equilibrium.
33
Fig. 4.10. Expected budget for the project, with and without considering sponsor collaborations.
34
5. DEVELOPMENT OF THE FEA MODEL
5.1. Brief introduction to the Finite Element Method and software.
Finite Element Analysis, or FEA, also known as Finite Element Method, or FEM, is a
computational method to numerically solve differential equations, with many applications
in the engineering field [Altair, 2015]. FEM can solve complex structural problems, such
as the one developed in this work, by separating the model into discrete elements in a
process known as meshing, resulting on a finite number of points where the solution can
be numerically found. The accuracy of the results will be dependant on the number and
quality of the elements in the mesh, the accuracy of the loads applied to the elements and
their properties, and the effectiveness of the numerical algorithm solving the problem,
directly related to the maximum residual value allowed for the solution.
FEA requires a complex, three-step suite of software, kindly provided by Altair Spain
as part of their sponsorship agreement with the Team:
• Pre-processor: Required to adapt and clean the geometry, mesh and introduce me-
chanical and material properties, loads and other boundary conditions. The software
used was Altair HyperMesh.
• Solver: The set of algorithms capable of solving the numerical problem. The soft-
ware used was Altair OptiStruct, a linear-static NASTRAN based commercial code.
• Post-processor: Necessary to obtain a visual representation of the results, which

can then be analyzed. The software used was Altair HyperView.
5.1.1. Mesh quality
Since the loads and load cases will be expanded upon on section 5.2, the mesh quality
will be discussed here. Mesh quality is dependant on the size, shape and connectivity of
the elements, and will affect the accuracy of the results obtained and the time needed for
the solver to compute a solution to the problem.
Mesh element size is defined by the user, and is the most relevant factor in the quality
of the analysis. As the discretization of the model improves, that is, the elements become
smaller and more numerous, the results of the analysis will become closer to reality. How-
ever, a delicate balance must be found, as increasing the number of elements will also
significantly increase the meshing and solving times, as the computational demand of the
model increases exponentially.
Element shape can be chosen by the user or left for the meshing algorithm to decide.
Mesh elements of different sizes are more appropriate under certain circumstances, al-
35
though successful meshing can be accomplished with different shapes in most cases. On
this project, four types of mesh elements, in 2- and 3-D, were used:
• 2D TRIA: with three edges and three nodes, the 2D TRIA element was used for
the more complex geometries, were quad elements would be unsuitable due to their
poor quality, and to ensure good mesh connectivity over long spans of 2D QUAD
elements, as determined by the mixed mesh algorithm.
• 2D QUAD: With four edges and four nodes, the 2D QUAD element was used for
the majority of the elements on the 2D surfaces of the model, following a quad-
dominated mixed mesh algorithm.
• 3D TRIA: With six edges and four nodes, the 3D TRIA element acts as a 2D TRIA
element in 3D solid meshes, covering complex geometries and filling out gaps left
by the 3D QUAD elements.
• 3D QUAD: With twelve edges and eight nodes, the 3D QUAD element constitutes
the majority of the elements on the 3D solid mesh, following a quad-dominated
tetramesh algorithm.
The shape of the elements also affects the quality of the model, with the geometric
parameters directly linked to the effectiveness of the solver:
• Maximum size: Defines the maximum size allowed for an element in the model.
• Minimum size: Defines the minimum size allowed for an element in the model.
• Warpage: Defines how out of plane with the surface a QUAD element is, and is
more relevant over curved surfaces. Is calculated by dividing a QUAD element into
two TRIA elements and compares the angle of the normal vectors to each TRIA,
for the two possible TRIA configurations.
• Skew: Defines the angle between the line connecting one node to the midpoint of
the opposite edge and the line connecting the midpoints of each remaining edge in
a TRIA element. Indicates how much difference there is from the element to an
ideal, equilateral TRIA element.
• Jacobian: Measures how close to ideal the element is, in a range from [-1] to [1],
with the element closer to ideal as it approaches the ends of the range, and less ideal
as the value approaches [0]. Since the solving process involves taking the inverse
of the Jacobian Matrix for the problem, failed elements will cause a unsolvable
mathematical error.
• Aspect Ratio: Measures the difference between the largest and smallest edges for
an element. Ideally, the aspect ratio should be 1.
36
• Taper: Measures the difference between the areas of four TRIA elements obtained
by dividing a QUAD element. Indicates the angular distortion of a QUAD element.
• Element Length: Most useful in dynamic or transient structural analysis, the ele-
ment length is used to check that there are no null-length elements on linear-static
analysis.
• Element Angle: Measures the maximum and minimum allowed angles between
edges for any element, which may cause implausible results.
The quality settings for the analysis vary with the dimensions of the problem. In this
case, the values are shown on figure 5.1.
Fig. 5.1. Mesh quality settings for the rim FEA model.
Finally, connectivity of the elements depends on mesh structure. Well organized

meshes with high-quality elements, such as the one shown on figure 5.2, from the final
FEA analysis for this project, will obtain more accurate results.
Fig. 5.2. Example of a high quality section of the mesh, on the change of curvature for the rim
spoke.
37
Mesh effects on the model will be discussed on section 6.4, with the mesh sensitivity
analysis of the model.
5.2. Loads and load cases.
Considering the dynamics of a race car and the conditions for assembly, four load cases
were selected, two relating to the maximum possible performance of the race car and
two relating to the mounting conditions for the tyre. The load calculations have been
simplified to the most restrictive case in order to reduce the total computational time for
analysis and leave more available time for design.
For operating conditions, maximum longitudinal and lateral acceleration where cho-
sen. In motorsport, forces are typically given as a value of times the acceleration due to
gravity, commonly known as G-forces and corresponding to a force per unit mass.
Maximum longitudinal acceleration happens during braking, as the braking force on
Formula Student prototypes is higher than their traction force on acceleration. Maximum
lateral acceleration happens mid-corner, as the car reaches the apex of the turn. Consid-
ering combined braking-and-cornering loads results in a less restrictive load case.
During the design phase, the target for braking force was set as 2 Gs and the target
cornering force was set at 2.5 Gs, with all components of the car designed and sized to
withstand, at least, that value.
For the tyre mounting conditions, two situations were considered, which will happen
every time a new tyre is mounted onto the rim. First, the jaws of the tyre mounting
machine will clamp down on the lips of the rim, while the steel head of the machine will
press the tyre sidewall past the lips. Following that, the tyre will be filled with air to a
pressure of around 3 bars (much higher than the typical operating pressures of the Hoosier
R25B tyres used by MAD Formula Team, which range between 0.8 to 1.2 bars).
5.2.1. Maximum Braking.
Translating the maximum longitudinal acceleration into a force is a simple matter. First,
during braking, a longitudinal load transfer happens, as a result of the action of Newton’s
First Law. Due to this longitudinal weight transfer, the front tyres will be submitted to
a larger force. Assuming that the car is rigid, and that the overall mass of the vehicle,
wheelbase and height of the center of gravity is known (which are design variables fixed
as part of the design process), as well as the target acceleration, equation 5.1 will show
the total weight transferred to the front tyres:
h
∆W = a · ·m (5.1)
w
Where:
38
∆W = Longitudinal load transfer, back to front, in N.
a = Longitudinal acceleration, in m/s2 .
h = Height of the center of gravity in m, a design variable.
w = Wheelbase of the car in m, a design variable.
m = Mass of the car in kg, a design variable.
With a static mass of the car, in kg, of 232 kg, plus the weight of the driver, estimated
during design to be 72 kg; a wheelbase of 1.545 m, and a maximum height for the center
of gravity of 0.320 m, the resulting load transfer is 1235.35 N.
Also knowing the weight distribution for the static car, which by design is 40:60 over
the front and rear axles respectively, the static weight on one of the front tyres and the
load transfer from the rear are added to obtain the total mass of each front corner. As
the acceleration is known, the force can be easily obtained again by applying Newton’s
Second Law, resulting in the total load for the rim, which will be transferred unto the road
through the tyre contact patch, shown on equation 5.2.
W = ∆W + WFC,static · a (5.2)
Where:
W = Longitudinal load on each front corner, in N.

∆W = Longitudinal load transfer, back to front, in N.
WFC,static = Static load on either front corner, in kg.
a = Longitudinal braking acceleration, in m/s2 .
Resulting on a total load for the corner of 2428.25 N during maximum braking.
5.2.2. Maximum Cornering.
Similarly, lateral weight transfer happens during lateral acceleration. In this case, in order
to obtain the lateral weight transfer, the known variables are the mass of the vehicle, the
track of the real axle (as there is no longitudinal weight transfer, the rear axle, carrying
60% of the total mass of the vehicle, is the most critical), the lateral acceleration and the
height of the center of gravity), which will be used in equation 5.3.
a·h
∆W = ·m (5.3)
T
Where
∆W = Lateral load transfer in N

a = Lateral acceleration in m/s2
h = Height of the Center of Gravity in m, a design variable.
T = Track of the car in m, a design variable
m = Mass of the car in kg, a design variable
39
With a wheel track of 1.170 m, and the same mass, mass distribution over the axles,
and height for the center of gravity, the resulting load transfer over the rear axle on cor-
nering results in 124.75 N of load transfer. Although this number may look low for such
a high lateral acceleration, the small dimensions of the car and extremely low center of
gravity cause very little load transfer during cornering.
Adding together the resulting lateral transfer plus the static load of each rear corner
will result in the mass for the most restrictive case. Considering the maximum lateral
acceleration, the resulting force on the higher-loaded rim is obtained as per equation 5.4.
W = ∆W + WRC,static · a (5.4)
Where:
W = Lateral load on each rear corner, in N.

∆W = Lateral load transfer, inside to outside corners, in N.
WRC,static = Static load on outer rear corner, in kg.
a = Lateral cornering acceleration, in m/s2 .
Resulting on a total load of 2362.45 N on the most critical cornering case.
5.2.3. Tyre Mounting.
Tyre mounting requires the rim to be held by steel jaws, usually three or four, over the
lower lip, whilst a steel head pushes the tyre past the upper lip. As the jaws and head
are small compared to the size of the rim, and separated on the lower lip by at least one
quarter, and sometimes even one third, of the rim circumference, whilst on the upper lip
the load is only applied at one point, we can consider the effect to be local, and simulate it
as a distributed load of 350N, the typical rating for tyre-mounting machines, over an area
of 625 mm2
5.2.4. Overpressure.
To seat the bead of the tyre on the rim, a sudden inflation is performed, increasing the
pressure of the tyres to around 3 bars before quickly dropping it to the target operation
pressure. This is the simplest load case, as a uniform pressure of 3 bars can be easily
applied to the inside surface between the tyre and the rim.
40
5.3. Load application.
5.3.1. Dynamic load application.
As usual in projects of this nature [Korntved et al., 2017], the load corresponding to the
dynamic working conditions of the rim, that is, the load corresponding to the longitudinal
and lateral acceleration of the car, is applied as a point force on a fictitious point, corre-
sponding to the point of contact between the tyre and the road, and distributed over the
inner bead of the rim through 1-Dimensional, non-stiff RBE3 connectors. Considering
the contact patch as a single point results in a more restrictive load condition than what
would happen in reality, where the deformation of the tyre would result in the load being
applied on a surface and more distributed over a larger section of the rim.
The point of application is located at a vertical distance of 228.6 mm from the center
of the rim, corresponding to the outer edge of a tyre with a nominal diameter of 18" (457.2
mm), in a parallel line with the axis of rotation of the tyre and 30 mm from the centerline
of the rim, which accounts for the offset from the rim mount to the hub to the centerline
of the rim. The hub was modeled as a no-displacement constraint, restricting the three
degrees of freedom in displacement over the X, Y, and Z axis, but allowing rotation,
applied over the edge of each of the fastener holes. Restricting rotation caused artifacts
and implausible results, with tensions above 1GPa over certain elements of the mesh,
with neighbouring elements in the 20-200 MPa range. Considering that restricting the
movement of each mounting hole separately restricts rotation for all intents and purposes,
the approximation can be considered valid.
The longitudinal load can be applied in either direction over the longitudinal X axis
of the car, as braking forces are equal on the right and left front tyres (corresponding each
to the positive and negative X direction on the model), whilst the lateral loads must be
applied in the positive Z direction. That is caused by the mounting of the rim, which is,
obviously, directional, and the fact that the loaded tyre on a corner is the outer tyre. Since
the default position on the rim on the model would be that of a left-hand side tyre, the
load would correspond to a right-hand turn, which following Newton’s Third Law would
cause a reaction force on the tyre to the left of the direction of travel for the car, that is,
positive over the Z-axis.
41
Fig. 5.3. Dynamic load application point on the rim, from the FEA model.
5.3.2. Mounting load application.
The mounting load application was applied differently for the overpressure and edge
loads. As discussed on section 5.2.3, a distributed load of 350 N was applied over a
random 25 by 25 mm section of the lip, with the results verifying that the effect of either
the grabbing claws or the mounting head is effectively local and can be simulated sepa-
rately. Since both lips have identical laminates and ply thicknesses, resulting in identical
area moments of inertia and mechanical properties of the material, simulations validating
just one load over one lip were considered sufficient.
42
Fig. 5.4. Mounting load application on the edge of the rim, from the FEA model.
The overpressure load was applied as a 3 bar pressure over the inside surface of the
rim, as discussed on section 5.2.4. Whilst linear-static analysis is not the most accurate
way to model a sudden increase and loss of pressure, initial results clearly showed negli-
gible displacements and Von Mises stresses one order of magnitude lower than the yield
strength of the material, which clearly indicates that there is a sufficient safety margin to
consider the transient-dynamic simulation of the overpressure unnecessary.
43
Fig. 5.5. Overpressure load application on the rim, from the FEA model.
5.4. Material Characterization.
For the design of the rim, three different materials have been used, each with different
mechanical properties and behaviour.
Aluminium was used for the laminated spacer at the center of the rim. The main
function of aluminium, in this case, is to act as a contact surface between the lugs and the
fragile carbon of the spokes, preventing damage to the spokes, and aligning the wheel and
hub. For this part, aerospace aluminium alloy 7075-T6 was selected, due to its excellent
mechanical properties, light weight and ease of machining, as well as its long history of
use by MADFT over the years, which results in carry-over knowledge and a know-how
base on its use beyond that of other engineering materials such as steel or other aluminium
alloys, such as 6063-T6.
For the pre-impregnated CFRP, three different materials were available for use, kindly
donated by Gurit Holding as part of their sponsorship agreement with MAD Formula
Team, namely XC305 and RC200T woven fiber cloths and UD200 unidirectional fiber
cloth. Whilst ideally the decision to use any of these available materials would be taken
on the basis of performance, the required use of RC200T and UD200 cloth for the man-
ufacturing of the monocoque left only discards and off-cuts for both materials, leaving
XC305 as the only viable option.
44
Finally, the spokes of the two-part rim model were designed for manufacturing using
short-strand CMCFRP. The two main benefits of this process are the ability to cheaply
manufacture complex geometries and the reduction in materials cost, as the short-strand
carbon fiber reinforcement is chopped from the off-cuts and discards of dry carbon fiber
cloth, including material from rolls that were discarded by the manufacturer due to pro-
duction issues with the weave. Since unused carbon fiber is uneconomical to recycle by
other methods, using the discarded material as reinforcement for CMCFRP is also more
environmentally friendly than manufacturing new material for other processes, such as
wet lay-up or infusion molding.
As part of the engineering process for the design of the rims, the unknown mechanical
properties of the CMCFRP material had to be ascertained. In close cooperation with our
MotoStudent colleagues, MOTO-MAQLAB-UC3M, four samples of known dimensions
were manufactured and submitted to traction tests until failure, registering the elongation
and force, from which stress-strain curves for each sample were plotted. The samples
were manufactured using EasyComposites BLACKSTUFF woven carbon twill, separated
by hand and chopped into 6 mm strands, and EasyComposites EL2 epoxy laminating resin
with AT30 FAST epoxy hardener agent. To ensure that the tests were not affected by fiber
direction, that is, that the behaviour of the material was truly quasi-isotropic, samples
were manufactured with a thickness of at least 15 mm in every dimension.
Unfortunately, due to human error, Sample 4 was unusable, as incorrect setup of the
elongation measuring apparatus caused implausible results. However, the remaining three
samples were enough to characterize the material as quasi-isotropic, with excellent linear
stress-strain behaviour and a Ultimate Tensile Strength coincident with the Yield Strength
of the material, as expected with carbon-reinforced composites. Results, shown graphed
on figures 5.6, 5.7 and 5.8, and mechanical properties, shown on table 4.1, were satisfac-
tory, as the material shows properties similar to some aluminium alloys but with lower
density. As with other parts of this project, time, budget and apparatus access limitations
were the reason for not performing more tests.
Once the properties for all materials were know, the model was updated to include
the relevant data. For the 7075-T6 aluminium alloy and the CMCFRP, the MAT1 ma-
terial model was used, corresponding to linear-elastic, isotropic materials, which is the
preferred approach for metallic engineering materials, as well as a good representation of
the data obtained during the physical testing of the CMCFRP samples. For the XC305
pre-impregnated CF, MAT8 was the material model used, corresponding to ortothropic
2D materials, a good fit for woven carbon fibre cloth.
45
SAMPLE MAX STRESS [MPa] MAX STRAIN [%] YOUNG’S MODULUS [GPa] POISSON RATIO SHEAR MODULUS [GPa] DENSITY (T/mm3 )
P1 328.2178 0.5653 58.05 0.084 26.79 1.44e-9
P2 242.4760739 0.292647604 83.61 0.104 37.86 1.49e-9
P3 215.5319731 0.291469271 74.32 0.091 34.06 1.41e-9
P4 359.2821759 1.929380208 Discarded Discarded Discarded Discarded
Average 286.3769978 0.287357344 71.99 0.093 32.90 1.45e-9
Table 5.1. MECHANICAL PROPERTIES FOR CMCFRP, AS PER

SAMPLES TESTED.
Fig. 5.6. Adjusted Stress vs. Strain curve for sample 1.
46
5.5. Fatigue.
Fatigue analysis of the CFRP rims was not performed. The reasoning behind this decision
is based on three arguments, to note:
• Fatigue analysis for composite materials requires study beyond the scope of this
project to build an accurate fatigue model for a specific laminate and set up. Given
the time restrictions for the project, the different laminates modelled and the re-
stricted access to the required apparatus for sample manufacturing and testing,
proper fatigue analysis could not be completed in time.
• Over the course of a single Formula Student competition, cars will drive around
30km over one week, depending on the number of runs allowed in the Practice
Area, which is unlikely to vary much. Even considering that a team attends all
competitions (except for those that overlap on the schedule) and that the team com-
pletes enough successful testing outings to pass the 500 km of running mark (an
exceptionally high distance for a Formula Student team on any season), the rims
will likely not suffer from any fatigue-related damage, as the running time, distance
and load cycles are very small.
• Finally, it is part of the realities of auto racing that most parts are designed to fail.
Oversizing components leads to additional weight, which, as shown, is a key factor
in reducing the performance of the car. Expected lifetime for many race car com-
ponents can be measured in hours, as opposed to normal, road-going cars where
lifetime and replacement cycles for components are measured in years or even in-
finite. While automotive manufacturers typically do not expect the end users to
47
replace their rims after every long drive, one of the advantages of racing is that re-
placing the rims after every competition for inspection is well within the mission
parameters.
48
6. DESIGN AND ANALYSIS.
6.1. Design process. CAD model.
The design process began by creating a basic rim design that could be easily edited to
test different spoke configurations. Since this is, broadly speaking, a weight optimization
problem, removing as much material as possible is the objective for the design phase
[Bhagwat, 2017].
In order to ensure compatibility with the current setup of the MADFT MFT02 race
car, so that testing sessions can be carried on during the design phase of the MFT03 race
car, the design of the rim must be of the same profile and offset (30 mm) as the current
BRAID alloy rims. However, using a central aluminium spacer will allow the Team to
manufacture rims with different offsets as required by design decisions, opening new
possibilities, on top of protecting the rim from damage from the wheel hub fasteners.
Diameter and width of the rim were decided for compatibility with the current tyre
selection employed by MADFT, the Hoosier 10" (254 mm) by 7" (177.8 mm) R25B and
R20B models. For the fasteners to the hub, four 12 mm holes need to be evenly spaced
over a circumference of 100 mm of diameter from the center of the wheel. Reception of
the centering component of the hub requires a 64 mm hole in the center-point of the rim.
Over the lip, one hole must be made for the insertion of the tyre inflation valve.
For the lips, initial design approach followed the Tire and Rim Association (TRA)
standards. However, considerations regarding the mold design, and specifically, the lami-
nation of some sections of the lip inner bead curvature and demoulding of the entire rim,
forced a reconsideration of the design. The TRA standards are oriented towards metallic
rims, which can be sand-casted, where demoulding is not an issue to consider, but for the
reusable molds for the CFRP rims, the standards were non-viable. A simplified version
of the BRAID rim cross-section was designed, losing some bead-locking features, which
will be substituted with the application of bead tape, a type of commercial adhesive tape
used to seat and lock the bead of tyres.
With those parameters, a CAD model of a base rim, with no spokes or features other
than the centering and fastener holes, was created using the Dassault Systèmes Solid-
Works CAD suite, kindly provided as part of their sponsorship agreement with MAD
Formula Team. From this base rim, different spoke profiles were cut out, with four, five,
six, eight and twelve spokes. Wheels with less than four spokes quickly become either
very heavy, as there is very little material removed to make a large difference in weight,
and wheels with more than twelve spokes quickly run into manufacturing problems, as the
spokes become very thin and slender, which makes demoulding hard and can also lead to
compromised structural capabilities. Some of these variants are shown in figures 6.1, 6.2
49
and 6.3. All models were submitted to analysis, in order to identify the best performing
spoke configuration in terms of weight and strength.
Fig. 6.1. Four-spoke rim variant.
Fig. 6.2. Five-spoke rim variant.
Fig. 6.3. Twelve-spoke rim variant.
50
All models were submitted to analysis, in order to identify the best performing spoke
configuration in terms of weight and strength.
6.2. Spoke design analysis.
Taking the base FEA model developed in section 5, the different designs for the spokes
were analyzed. To save time and computational costs, instead of developing a full model
for each rim design, a comparative study was performed. Rims with all different spoke
configurations were characterized as 2D shell surfaces with a set thickness of 4 mm,
and meshed using the same parameters. To keep mesh quality similar, the meshing and
element clean up were performed using Altair Hypermesh automated tools, preventing
human interaction from affecting the results. A material was then assigned to the rims
in order to carry on the study. Considering the fact that this is a comparative study,
any random material properties could have been assigned. However, in the interest of
being able to properly understand the results, aluminium alloy 7075-T6 was selected, as
it was already one of the materials in the model, and also produced results that could be
interpreted within the background of the current rim setup for MADFT, using aluminium
BRAID rims. The load applied was oversized, in order to produce large enough results
for the differences to become easily noticeable. Lateral and longitudinal loads of 35 kN
were applied, resulting in displacements in the tens of millimeters range and Von Mises
stresses up to the gigapascal range, with the weight of the rims around the 2.5 kg mark
for every design. Large displacements and stresses were par for the course, as no plastic
or failure behaviour of the material was modelled so as to force large results were the
differences could easily be seen. An example, showing the comparative analysis of the
four-spoke rim, can be seen on figures 6.4, 6.5, 6.6, 6.7, 6.8, and 6.9. Only lateral and
longitudinal loads were used, as they are the performance-oriented load case, meaning
that the on-track performance of the rim is directly correlated to them, and because for all
spoke designs the rim lip and bucket remain the same, which are the parts of the rim that
will withstand the brunt of the mounting and overpressure loads.
51
Fig. 6.4. Lateral load application on the 4-spoke rim variant.
52
Fig. 6.5. Von Mises stresses, in MPa, for the 4-spoke rim variant submitted to lateral load.
Fig. 6.6. Displacement, in mm, for the 4-spoke rim variant submitted to lateral load.
53
Fig. 6.7. Longitudinal load application on the 4-spoke rim variant.
54
Fig. 6.8. Von Mises stresses, in MPa, for the 4-spoke rim variant submitted to longitudinal load.
Fig. 6.9. Displacement, in mm, for the 4-spoke rim variant submitted to longitudinal load.
55
The results, shown on table 6.1, indicate that rims with numbers of spokes multiples
of one another have very similar results. This is to be expected, as the load path for the
lateral and longitudinal accelerations is through the edges of the spoke, not the center.
Since doubling, in our case, the spokes effectively means hollowing out the center area
of the spoke to create two more where the edges would be, the load paths remain mostly
undisturbed and offer similar results. Increasing the number of spokes has very little
effect on the weight, which for the 4-, 5-, 6- and 8-spoke design varies less than 20 g,
but it does affect the Von Mises stresses and displacements in opposite directions: more
spokes reduces displacement but increases the internal stresses on the material. The most
logical explanation for this effect is that a larger number of spokes offers more bracing to
the rim, limiting displacement, but the thinner spokes, with smaller cross-sections, suffer
more stresses in turn.
LATERAL LOAD LONGITUDINAL LOAD
CONFIGURATION MASS [kg] STRESS, VON MISES [MPa] DISPLACEMENT [mm] STRESS, VON MISES [MPa] DISPLACEMENT [mm]
4 SPOKE 2.269 1703 26.67 488 4.609
5 SPOKE 2.258 1895 23.21 563 4.210
6 SPOKE 2.271 1982 22.89 589 4.017
8 SPOKE 2.265 1721 22.29 502 4.091
12 SPOKE 2.468 1978 22.26 611 4.009
Table 6.1. MAXIMUM VON MISES STRESS AND ABSOLUTE

DISPLACEMENT FOR EACH SPOKE CONFIGURATION AND
DYNAMIC LOAD CASE.
As shown by the results on table 6.1, the best performing rims were either the four-
spoke or eight-spoke designs, with very little difference between the two in terms of
weight, displacement or Von Mises stress, despite forcing large scales of results. To
break the tie, a final consideration had to be taken into account: most of the budget for the
design and manufacturing of MADFT’s race car comes from sponsors, who are attracted
to the project for a variety of reasons beyond the control of the Team. However, designing
an attractive looking car may help those on the fence about participating on the project
make a decision, on top of offering a better image on social media and events. A very
simple comparison between the two designs was carried out, asking 17 team members
to rank both rims between 1 (strongly dislike) and 5 (strongly like). The results, shown
on figure 6.10, showed not only strong support for the eight-spoke rim design, but also a
strong dislike for the four-spoke rim.
56
Fig. 6.10. Preference in terms of overall aesthetics by 17 members of MAD Formula Team.
Given the data from the simulations and the Team’s clear preference for the eight-
spoke rim, the simple 2D shell design for the comparative analysis was refined as a 3D
solid for final analysis and mold design.
6.3. Two-part vs. Single-part rim.
Two manufacturing options were considered: the rims could be built as a single part,
including the bucket and spokes in a single mold, and laminating the entire assembly to-
gether, or as two parts, manufacturing the spokes and bucket separately before bonding
them together as an assembly. The key differences between both manufacturing methods
are the final strength and stiffness of the rim and the overall cost of manufacturing. Fur-
thermore, single-part manufacturing makes the use of CMCFRP very difficult, as the mold
requirements for pre-impregnated CFRP are incompatible with the mold requirements for
CMCFRP, making the use of the later material, which is a key part of achieving a low
per-unit cost, non-viable.
Both options were assessed through FEM for structural integrity, beginning with the
single-part. Considering that the CMCFRP spokes were modeled as a three-dimensional
solid with quasi-isotropic properties, while pre-impregnated CFRP was considered as a
two-dimensional ortothropic material, the newly designed 3D spokes had to be converted
again into a 2D shell surface, using SolidWorks’ surface tools to prevent loss of detail.
The completed shell was then submitted to the loads calculated in section 5.3, obtaining
the results shown on table 6.2.
The two-part rim was modeled with a separate bucket and spokes, with the aluminium
spacer added to the model separately from the spokes and then joined using perfectly
57
stiff RBE2 1D elements. The bonding process between the two parts, which calls for
structural two-component epoxy-based adhesive and rivet pins, was modelled as a perfect,
no gap bond through the use of the Hypermesh Freeze contact condition, and the rivet pins
modelled as 1D S355 steel beams with a 3 mm diameter. The model was submitted to the
loads calculated in section 5.3, obtaining the results shown on table 6.2.

ONE PART TWO PART ONE PART TWO PART
MAX STRESS [MPa] 1721 1812 502 523
MAX DISPLACEMENT [mm] 22.29 22.45 4.091 4.112
Table 6.2. DIFFERENCES IN STRESS AND DISPLACEMENT FOR

ONE- AND TWO- PART RIMS.
Whilst single-part manufacturing, as expected, produces a rim that is stiffer and has
lower overall Von Mises stresses than the two-part manufactured rim, the difference, as
can be seen on Table 6.2, is very small. However, the difference in expected manufactur-
ing costs, as seen on 4.10, clearly indicates that two-part manufacturing is the best option
in line with the stated objectives for this project, that is, low-cost manufacturing.
6.4. Final design and FEA model.
Having decided that the most suitable solution was to use a two-part rim, eight-spoke rim,
the original, plain design was improved upon. The 2D spoke surface was substituted for
an 3D design, featuring arched radii and a more aggressive look, more in line with the
overall aesthetics of the race car. The lip cross-section was fully defined, to allow for
parametric modifications should they be needed in the future, and the offset of the spacer
was given as a variable for future suspension set up.
58
Fig. 6.11. Parametric dimensions of the lip.
Fig. 6.12. Final model of the CFRP rim.
With the CAD design finished, the final FEA model was developed. The two key
parts of the rim, the bucket and the spokes, were modeled as a 2D surface and a 3D solid,
respectively. The thickness of the spokes was parameterized to allow for rapid analy-
sis of different thicknesses, while laminate plies in the 0º/90º and 45º/-45º configuration
were set up. The laminate was designed as a symmetrical laminate, in order to avoid un-
desirable behaviours such as shear-twisting or extensional-building coupling behaviours,
which affect thin laminates and can cause catastrophic failure under loaded conditions.
On the model, as explained on sections 5.2.1, 5.2.2, 5.2.3, and 5.2.4, the four con-
sidered loads were applied, with the dynamic loads on a point corresponding to the tyre
59
contact patch and the static loads applied locally to the areas of interest. The hub con-
straints remained constant for all load cases.
A mesh sensitivity analysis was carried out to assess the mesh-dependency of the
results and also optimize the computational time for the model ensuring convergence of
the results. The meshing criteria, as discussed on section 5.1.1, was selected as mixed,
with element size ranging from 7 to 1 mm as shown on figures 6.13 and 6.14. Parameters
of interests on the mesh sensitivity analysis were maximum displacement, maximum Von
Mises stress, and meshing and computation times for the model and solution, respectively.
Fig. 6.13. Element size vs. Displacement (both in mm) for the mesh sensitivity analysis.
Fig. 6.14. Element size vs. Other relevant variables for the mesh sensitivity analysis.
From the results of the mesh sensitivity analysis, the 2 mm mesh size was selected, as
the results on figure 6.13 and 6.14 show a convergence of results at around that mesh size.
Furthermore, the increase in computational power required, depending on the number of
degrees of freedom of the model (which, in turn, depends on the number of elements),
grows steeper as the mesh size decreases.
60
With all the model variables defined, iterations were made to find the optimal laminate
for the bucket and optimal thickness for the spokes. Considering that the wheels are a
safety-critical component, where failure may cause damage to both the car and driver, a
relatively high safety factor of 1.65 was chosen. This may result shocking to people used
to structural safety factors above 2 or even 3, but it is within the expected parameters for
race car design, were components such as the monocoque may have a safety factor for
the most critical scenarios of 1.2. A first laminate and spoke thickness, of just two plies
and 0.5 mm respectively, was analyzed, and the thickness gradually incremented until it
fulfilled the safety failure criteria.
For the failure criteria of the CMCFRP spokes and aluminium spacer, the Von Mises
stress criteria was applied, with a limit of 286.77 MPa for the CMCFRP, as per the data
on table 5.1, and a limit of 450MPa for the 7075-T6 aluminium alloy.
Failure criteria for the XC305 laminate CFRP was defined by the maximum strain
criterion, as there is experimental verification [Lee and Roh, 2015] that the criterion is
suitable for isotropic laminate composite materials. The criterion defines failure at the
point where the deformation in local coordinates for a single ply of the material are su-
perior to the maximum deformation in shear, longitudinal or transverse directions to the
fibers of the ply, and is represented by a single value, the Composite ply failure index. For
a value of 1 or larger, the composite laminate will fail.
Results for the iterations showed that the mounting loads did not cause failure in all
but the initial iteration, and that the aluminium spacer, for which the dimensions are fixed
by suspension parameters and cannot be changed, would not fail under any conditions.
In table 6.3, results showing the maximum Von Mises stresses for the dynamic loads, as
well as the composite ply failure index and weight for the rim, are shown.
LAMINATE PLIES SPOKE THICKNESS [mm] MAX VM STRESS [MPa] MAX DISPLACEMENT [mm] MAX VM STRESS [MPa] MAX DISPLACEMENT [mm] COMPOSITE PLY FAILURE INDEX, WORST RIM WEIGHT [g]
0/90, 0/90 0.5 302.0 1.621 288.6 1.992 1.230 512.22
0/90, 45/-45, 45/-45, 0/90 1.5 221.8 1.002 205.2 1.007 1.009 624.4
0/90, 45/-45, 0/90, 45/-45, 0/90 2.5 192.9 0.819 189.1 0.911 0.782 762.4
0/90, 45/-45, 0/90, 45/-45, 0/90, 45/-45, 0/90 3 177.2 0.633 174.5 0.812 0.612 898.2
0/90, 45/-45, 0/90, 45/-45, 45/-45, 0/90, 45/-45, 0/90 3.2 170.9 0.489 167.5 0.717 0.408 964.9
Table 6.3. VON MISES STRESSES, DISPLACEMENTS AND

COMPOSITE FAILURE INDEX FOR EACH LAMINATE AND
SPOKE THICKNESS ITERATION, PLUS THE WEIGHT FOR THE
CORRESPONDING RIM.
6.5. Mold design and manufacturing.
Molds were designed as two sets of two-part molds, one for the spokes and one for the
bucket. The molds are designed for 3D printing [Thuening, 2018], requiring a minimum
bed size of 460x460 mm2 , although other material options, as shown on the budget in
figure 4.10, were also considered. A 1:2 scale model of the molds was manufactured to
check the viability and the process.
61
Fig. 6.15. 3D printing of one of the mold components.
The bucket mold is designed to be laminated on the outer surface, which is also
known as a male mold configuration, with the vacuum bag required for curing the pre-
impregnated CFRP draped over the naked carbon fiber. The mold sits on the inside of
the rim and separates along the rotation axis of the wheel at the outside lip, allowing the
finished part to be pulled up and demolded once cured, as shown on figure 6.16.
Fig. 6.16. Mold for the rim bucket, separated into its two components.
62
Fig. 6.17. Mold for the rim bucket, with both components assembled together.
63
Fig. 6.18. Mold for the rim bucket, joined with threaded rods before lamination. Small damage to
the upper lip was easily repaired.
On the other hand, the spoke mold is a compression mold, as befits CMCFRP, with a
cavity that is filled with the appropriate mixture of chopped carbon fiber tow and epoxy
resin, and then pressed with a plunger mold to cure the material under pressure, as shown
on figure 6.21. About 80 kg of pressure are required to seat and cure the part properly,
easily obtained with traditional, crank-operated bench vices.
64
Fig. 6.19. Molds for the rim spokes, separated.
Fig. 6.20. Male (plunger) mold for the spokes.
65
Fig. 6.21. Molds for the rim spokes, assembled together.
Once both parts have been manufactured, eight drilled holes are made on each, fol-
lowing small bumps made on the surface of the molds that serve as guides to ensure even
spacing, and the contact surfaces for the bonding are roughed up and cleaned to increase
surface area. Eight rivets are introduced on the holes in the bucket, with a washer to in-
crease load distribution area over the composite material, before applying the adhesive
to both parts and joining them, with a short cure of two hours. After the bonding of the
adhesive is confirmed, the two parts are riveted together placing washers on the spoke. A
final hole is drilled to insert the air pressure valve, and the entire assembly is sent to be
balanced, as with any other rim.
6.6. Analysis and results.
The FEM analysis shows a very capable, meaning lightweight and strong, rim. With
displacements below one millimeter, and a 1.65 coefficient of safety in terms of stress, the
rim can be safely used as part of a racecar. The results in figures 6.22 and 6.23, showing
Von Mises stresses over the rim for lateral and longitudinal accelerations, are consistent
with the expected stress distributions for the rim, as well as results from other similar
projects [Kandukuri et al., 2022].
66
Fig. 6.22. Final Von Mises stresses, in MPa, for the lateral acceleration load case.
Fig. 6.23. Final Von Misses stresses, in MPa, for the longitudinal acceleration load case.
Displacements under dynamic load are very small and also in line with the results
obtained by other similar studies, and show that the rim and tyre are very unlikely to
separate under normal load conditions, as shown on figures 6.24 and 6.25.
67
Fig. 6.24. Final displacements, in mm, for the lateral acceleration load case.
Fig. 6.25. Final displacements, in mm, for the longitudinal acceleration load case.
Finally, the composite failure criterion is fullfilled in both dynamic cases, with a higher
factor of safety than the CMCFRP parts, which is also to be expected, as the mechanical
properties of the pre-impregnated twill are superior to those of chopped carbon. As shown
68
on figures 6.26 and 6.27, in neither case the composite failure is cause for concern.
Fig. 6.26. Final composite ply failure index for the lateral acceleration load case.
Fig. 6.27. Final composite ply failure index for the lateral acceleration load case.
69
Finally, the results in terms of mass are satisfactory, with a final weight of 964 g as
calculated by the FEM suite, as shown on figure 6.28.
Fig. 6.28. Final weight of the model in tons, as per the FEA suite.
With regards to the tyre assembly loads, as per sections 5.2.3 and 5.2.4, a tyre can be
assembled unto the rim without problem. During tyre mounting and inflation, the Von
Mises stresses, displacements and composite failure criterion are fulfilled with more than
enough margin of safety to confidently state that they are a non-issue, as shown on figures
6.29, 6.30, 6.31, 6.32, 6.33, and 6.34.
70
Fig. 6.29. Final displacement, in mm, for the tyre mounting load case.
Fig. 6.30. Final displacement, in mm, for the overpressure load case.
71
Fig. 6.31. Final Von Misses stresses, in MPa, for the tyre mounting load case.
Fig. 6.32. Final Von Misses stresses, in MPa, for the overpressure load case.
72
Fig. 6.33. Final composite ply failure index for the tyre mounting load case.
Fig. 6.34. Final composite ply failure index for the overpressure load case.
73
7. CONCLUSIONS.
7.1. Design conclusions.
The analysis of the rim shows a fulfillment of all initial criteria for the project: light
weight, good stiffness, and able to withstand the required operating and assembly loads,
as well as introducing new manufacturing techniques and materials to the Formula Student
scene.
As shown on figure 6.28, the rim design weights 954 g, which is well within the 1.25
kg target as discussed on section 4.4. Since there is very little difference in terms of perfor-
mance against the lightest possible CFRP rim, which presents a much bigger challenge in
terms of optimization and is very unlikely to be within the budget for MADFT, as shown
on figure 4.10, any further optimization, whilst certainly possible, would be unproductive.
The results in terms of stiffness, or, more accurately, reduction of displacement, are
also wholly satisfactory, as the maximum displacement of the rim is below the maximum
expected displacement of the tyre under dynamic load. Formula Student tyres, as with
most racing tyres, are very soft and pliable at working temperature and under load, and
the contact patch can move up to a few millimeters in either direction when loaded. Since
the movement of the tyre is at least one order of magnitude higher than the movement of
the rim, it is very unlikely that the tyre can become unseated during a dynamic event.
The safety requirements are also fulfilled, with all failure criteria within the 1.65 fac-
tor of safety stated, and in some cases even over 1.80 or 2.00. Although, again, this may
be a surprising safety factor to engineers used to other structural projects, car racing is
inherently a dangerous sport, and much lower safety factors are allowed and even encour-
aged.
Lastly, the introduction of a new manufacturing method and material into Formula
Student was received with moderate success. Although the rims were not manufactured
before the 2022 competition season began, some parts were developed using CMCFRP
as parallel side projects during the season, with some lessons learned during the process
used to refine the FEM model for the rim, and feeding back important data, such as the
mechanical properties of CMCFRP, to the projects.
Therefore, the process can be considered, at least, a success, with the conclusions to
the viability of the low-cost manufacturing discussed on section 7.2.
74
7.2. Viability of low cost manufacturing.
Apart of the design of the rim, which, as seen on section 7.1, can be considered a success,
the other key objective for this project was to adapt the rim for low cost manufacturing.
Unexpected time and access constraints during the season, as well as the expected
budgetary constraints, meant that the final step for the development of the rims, manu-
facturing of an actual, real scale prototype for testing, could not be achieved. However,
a 1:2 scale of the rim bucket was manufactured, as well as one set of molds for the 1:2
scale spokes. As expected, the rim manufacturing process found some issues, which, as
described on section 8, can be fixed in future iterations.
At this point, the expectation is that low-cost manufacturing is a real possibility for
the next season, but with an understanding that it can only be achieved if the sponsors
agree to at least the same level of support as in previous seasons.
Without sponsor support, it is more economical to either find commercial CFRP rims
or use a lightweight magnesium alloy and invest the difference in cost into other areas in
order to balance the performance vs. cost of the car.
7.3. Environmental Impact.
Discussing the environmental impact of a race car is, at best, complicated. Whilst the
recycled nature of CMCFRP can be a net positive to the environment, the final purpose
of the race car itself is extremely contaminating, negating any positive effect that the
recycling of carbon fiber (an extremely polluting material at the end-of-life cycle, in any
case) may have had.
As shown on figures 4.6 and 4.7, there is also a benefit in fuel efficiency, which trans-
lates into a smaller carbon footprint for the entire team. However, the actual running time
of the car is so small that the results are almost negligible.
Despite the best intentions of the author, car racing has a large environmental impact
that cannot be easily mitigated by redesigning components or developing new ones, and
instead must rely on external carbon offset processes to reduce the overall carbon footprint
of the team. It is unclear if the development and manufacturing of the rims covered in this
project would be a net carbon offset with regards to using CFRP rims if the Team had
to buy them from external suppliers, such as BRAID, and have them delivered, without
additional data on their operations, manufacturing and logistics.
75
8. FUTURE DEVELOPMENTS.
The most obvious future development would be, of course, manufacturing the rims.
Due to time and budget constraints, it was not a priority for the Team to complete the
manufacturing during the 2021-2022 design and manufacturing cycle.
However, a 1:2 scale model for the rim, compatible with 5" karting tyres, was de-
signed, with molds manufactured in 3D printed PLA, in order to perform static and dy-
namic testing on a scale model. However, manufacturing issues with the car during the
season left very little time to complete the scale model, with only the bucket of the rim
laminated in time. As a first test-bed of the manufacturing process, some problems were
identified:
• PLA is not naturally self-releasing, like other 3D printing materials such as PETG,
which does not bond with the epoxy matrix, and significantly complicated the ex-
traction of the part from its mould, requiring the partial destruction of the mold
despite having used a commercial mold-release agent. This is easily solvable by
either using self-releasing materials, such as the aforementioned PETG, or moving
over to aluminium molds, which can be better coated with the mold-release agent,
if the available budget allows for it.
• Overlap between the different laminates on the bucket is both difficult to manufac-
ture and difficult to accurately model on FEA. Once the entire rim is manufactured
and assembled, static load tests on the overlap areas will have to be carried out in
order to validate the model. Special attention during manufacturing will be required
to minimize the risk of delamination, as the prepreg sheets may not properly bond
on the overlap, and Non-Destructive Testing (NDT) will be required to verify each
rim before final assembly using ultrasound or X-ray imaging.
• Although the spokes were not manufactured, a problem was identified with the
mold before manufacturing. The demoulding angle for the edges of the spokes was
too steep, which would seriously complicate the manufacturing of the part, probably
requiring destruction of the molds. This problem can be easily fixed in the future,
and is no cause for concern.
• Without performing real-world tests, it is unlikely that the bonding process between
the rim and the spokes, using structural adhesive and rivets, can be accurately mod-
eled. The current model supposes perfect bonding, which could be an accurate
representation if a traction test pulling the spokes out of the rim results in forces at
least one order of magnitude higher than the maximum service load required to sep-
arate both components. However, better modelling solutions will have to be found
if that is not the case, such as modelling a bonding patch of sorts with mechanical
76
properties equal to that of the cured adhesive, and 1D beam elements with the di-
mensions for the rivets, instead of the current condition. This development requires
both real-world testing and high computational resources.
Given the high increase in performance associated to the use of CFRP rims, parametriza-
tion of the design would allow other teams to receive a substantial boost to their per-
formance. Although it may result shocking to some that helping rival teams, who are,
effectively, our direct competitors, is part of the roadmap of future developments, one
of the key philosophies in Formula Student is sharing knowledge and know-how. By
parametrizing some of the main design variables for the rim, namely diameter, width, and
offset, which can be easily done using the Formula tool on the SolidWorks software suite,
we can democratize access to low-cost CFRP rims for low-budget teams, helping them
break into the top rankings and taking the spotlight away from highly-backed high budget
teams.
Finally, CMCFRP is a process that presents itself well towards automatization. Deriv-
ing a commercial patent for the design of the rim, molds and tooling required for series
manufacturing would not be out of reach for the current capabilities of MAD Formula
Team, which would provide a new source of funding for the Team and make the rim
available for widespread use.
77
BIBLIOGRAPHY
Altair. (2015). Practical aspects of finite element simulation. Altair.

Bhagwat, P. (2017). Design and analysis of a 10 inch carbon fiber wheel for a formula
sae racecar (Graduate Thesis, University of Texas Arlington).
BRAID. (2022). Sturace carbon fiber rim. Retrieved September 6, 2022, from https : / /
www.braid-wheels.com/en/sturace-carbon-fiber/P-1067/C-14684?L2VuL2Zvcm11bGEtc3R1ZGVu
Rules 2022. (2021). https://www.formulastudent.de/fileadmin/user_upload/all/2022/
rules/FS-Rules_2022_v1.0.pdf
Kandukuri, S. Y., Pai, A., & Manikandan, M. (2022). Scope of carbon fibre-reinforced
polymer wheel rims for formula student racecars: A finite element analytical ap-
proach. Journal of The Institution of Engineers (India): Series C, volume 103,
939–948. https://doi.org/https://doi.org/10.1007/s40032-022-00808-w
Korntved, K., Hersboll, J., Lauritsen, M., Bording, K., & Leto, H. (2017). Design and
analysis of a formula student carbon fibre rim (THE 5TH STUDENT SYMPO-
SIUM ON MECHANICAL AND MANUFACTURING ENGINEERING).
Kruse, P., Wathen, H. J., Hyde, J. J., Dobben, N. D., & Blake, J. D. (2021). Design and
analysis of one-piece 10 inch carbon fiber wheels for zips racing zr20 formula sae
racecar (University of Akron).
Lee, S.-Y., & Roh, J.-H. (2015). Two-dimensional strain-based interactive failure theory
for multidirectional composite laminates. Composites Part B: Engineering, 69,
69–75. https://doi.org/https://doi.org/10.1016/j.compositesb.2014.09.032
Making 280mph capable carbon fiber wheels. (2014). https://www.youtube.com/watch?
v=PGGiuaQwcd8
ROTOBOX. (2022). Rotobox carbon fiber rims. Retrieved September 6, 2022, from https:
//www.rotobox-wheels.com/en
Seward, D. (2014). Race car design. Red Globe Press.
Thuening, S. (2018). Manufacturing a composite wheel prototype using 3d printed molds
(Graduate Thesis, Minessota State University).
Walther, H. (2016). Development of a lightweight laminated composite wheel for formula
sae race vehicles (Master Thesis, University of Kansas).
78
APPENDIX
CERTIFICATES AND DATASHEETS FOR THE REQUIRED COMPOSITE MA-

TERIALS
Certificate No:
TAK000015Z
This is to certify:
That the Biaxial (±45°) Carbon Fibre Prepreg
with type designation(s)

SE 84LV XC - Series
Issued to
Gurit (UK) Ltd.
Newport Isle Of Wight, United Kingdom
is found to comply with

DNV GL class programme DNVGL-CP-0431 – Type approval – Prepreg materials
Application :
Manufacturing of FRP laminates.

Issued at Hamburg on 2018-03-05

for DNV GL
This Certificate is valid until 2023-03-04.
DNV GL local station: Hamburg Materials & Welding
Approval Engineer: Guido Michalek Thorsten Lohmann

Head of Section
This Certificate is subject to terms and conditions overleaf. Any significant change in design or construction may render this Certificate invalid.
The validity date relates to the Type Approval Certificate and not to the approval of equipment/systems installed.
Form code: TA 251 Revision: 2016-12 www.dnvgl.com Page 1 of 2
© DNV GL 2014. DNV GL and the Horizon Graphic are trademarks of DNV GL AS.
Job Id: 262.1-026924-1
Certificate No: TAK000015Z
Product description
Biaxial (±45°) hot-melt epoxy carbon fibre prepreg with curing temperatures between 80°C and 120°C
and curing times between 1 hour and 12 hours (depending on curing temperature).
Approved variants
- SE84 LV XC305 - SE84 LV XC411 - SE84 LV XC611
Approved properties
Fibre Area Weight: 290 – 640 g/m²
Resin Content: 35 – 45 %
Fibre Type: High Tenacity (HT)1)
HT is the DNV GL definition according to DNV GL Class Programme CP-0096 and is comparable to
1)
Gurit’s HEC definition.
Type Approval documentation

- Technical Data Sheet
- Test Report No. 6828 issued by the DNV GL approeved (Approval No. AOSS0000DHK) Testing
Laboratory of Gurit (UK).
- Quality documentation
Assessed production sites

Gurit (UK) Ltd. Gurit (Spain) Ltd.
St. Cross Business Park Polígono Industrial Romica, Calle 4
Newport PO30 5WU Parcela 11C, Albacete 02080
United Kingdom Spain
Periodical assessment
A production site with a valid Approval of Manufacturer (AoM) certificate for material in question is
exempted from the obligation concerning retention and renewal assessments.
For manufacturer without a valid AoM a periodical assessment after 2.5 years and at renewal after 5 years
is required.
Remarks
None
END OF CERTIFICATE
Form code: TA 251 Revision: 2016-12 www.dnvgl.com Page 2 of 2

General Datasheet
SE 84LV
LOW TEMPERATURE CURE EPOXY PREPREG
¬ Versatile, high-strength prepreg system
¬ Curable at temperatures as low as 80°C (176°F)
¬ Can be processed with vacuum-only processing
¬ Excellent tack
¬ Low Viscosity – Ideal for use with heavy fibre weights
¬ Lloyd’s Register and DNV-GL Certified Formats Available
INTRODUCTION
SE 84LV is an exceptionally versatile hot-melt, epoxy prepreg. It can be cured at temperature as low as 80°C (176°F), or
can be used for faster moulding of components at 120°C (248°F). This is achieved with an extremely good outlife of up to 8
weeks at 18-22°C (64-72°F). It is a toughened system, and offers excellent mechanical properties on a wide variety of
reinforcing fabrics and fibres.
SE 84LV is commonly used in vacuum bagging, press-moulding, autoclave

and other pressure moulding processes.
SE 84LV is a very low viscosity system used with heavy fibre weights where
low-flow processing conditions (vacuum bag pressure and minimum cure
temperature), are likely to be used. With its high compressive strength it is
widely used in large heavily loaded components, such as yacht hulls, and
spars. It has been selected for use by various America’s Cup syndicates and
boats racing in the Volvo Ocean Race.
SE 84LV is widely used in sandwich structures with honeycomb, foam and

balsa cores, primarily with the toughened SA 80 Adhesive Film.
PDS-SE 84LV-27-0421 1
PROCESSING NOTES - GENERAL PROCESSING NOTES - CURING
PREPARATION CURE CYCLES
When preparing the lay-up the prepreg should be removed from the For a good balance of composite properties, it is
freezer and allowed to thaw in a sealed bag. This may take 6 to 24 recommended that the laminate is cured at 80°C (176°F) for a
hours depending on roll size. This prevents atmospheric moisture minimum of 12 hours. A laminate may be cured in two stages -
from condensing on the prepreg which may cause voiding on cure. if, for example, making a cored component. However in a two
The mould surface should be release coated and must have been stage cure, a minimum of 4 hours at 85°C (185°F) or 5 hours at
tested for vacuum integrity prior to lay-up. 80°C (176°F) is recommended before debagging a skin, and it
must be ensured that this skin is cured for the equivalent of at
LAYING-UP least 10 hours at 85°C (185°F) or 12 hours at 80°C (176°F)
The following procedure is recommended for preparing vacuum before going into service.
cured laminates.
SE 84LV may be cured at higher temperatures for a shorter
1. Place the lay-up on a tool or caul sheet which has been treated time. At a cure temperature of 100°C (212°F) cure can be
with a release agent or film. Insert a thermocouple into the lay-up achieved in 3 hours or at 120°C (250°F) cure can be achieved
near the centre ply of the thickest edge section, outside the net trim in 1 hour.
line. A separate prepreg nylon peel ply is available for covering a
mould tool prior to lay-up in order to leave a clean, textured surface It is not recommended to cure SE 84LV under vacuum
for subsequent bonding. pressures of less than 85%. If a ramp rate of less than
0.3°C/min (0.5°F/min) is used, users should satisfy themselves
2. Apply a peel ply to the surface of the lay-up. Note that for good that this allows adequate flow.
secondary bonding of a peel-plied surface of an SE 84LV prepreg
laminate, a nylon peel ply, such as Gurit’s Stitch Ply A, is strongly
recommended. This is particularly important where the cure
temperatures are in excess of 90°C (194°F). Cover the peel ply
entirely with a perforated release film. Normally, no edge resin
bleeder system is used. For thin sections, Gurit WL3600P90 grade
release film are recommended, while for sections of 4mm and
above, Gurit WL3600P release film is also suitable. With WL3600P
the amount of resin bled away is controlled by the number of dry
plies of resin bleeder cloth placed over the perforated release film.
3. Install a vacuum bag by standard techniques. Insert at least two CURING AT 80°C (176°F)
vacuum stems through the bag connecting one to the vacuum When curing at 80°C (176°F) it is important to ensure the
source and the other, at a point on the part furthest from the source, temperature is monitored off the trailing thermocouple. 80°C
to a calibrated vacuum gauge. Position part in the oven or autoclave (176°F) should be treated as the minimum cure temperature
and draw vacuum to check for bag or system leaks. for SE 84LV; 70-75°C (158-167°F) will not generate adequate
mechanical properties.
4. Commence the heat-up cycle, typically between 0.3°C(0.5°F)/min
and 2°C(3.6°F)/min to the final cure temperature. At 85°C (185°F),
the temperature should be held up for 10 hours. Faster cures may THIN LAMINATES
be obtained at elevated temperatures, e.g. 6 hours at 90°C (194°F), When using very thin laminates (e.g. with a total laminate fibre
3 hours at 100°C (212°F) or 1 hour at 120°C (248°F). All weight of less than 300gm2), care needs to be taken to avoid
temperatures measured by the previously installed thermocouple. extracting excessive amounts of resin during the cure process.
When curing at 80°C (176°F) a minimum of 12 hours is To avoid this, a microporous release film can be used, and for
recommended. Vacuum should be maintained as high as possible, particularly critical components, a prepreg peel ply should be
with a minimum of 85% throughout the cure cycle. used.
5. Upon completion of cure, turn off heat and cool until part
temperature has fallen below 60°C (140°F). When fully cooled, the
part may be debagged, trimmed and machined as necessary. A
post-cure is not required.
CORE BONDING
This product can be used in conjunction with typical core materials. Representative test panels should be made to ensure that the laminate
construction, curing method and other variables allow full filling of any cuts or slits in the foam. The cure cycles given in this datasheet are
for typical monolithic flat panels and may not be appropriate for sandwich panels.
When using Nomex™ or aluminium honeycombs, the separate SA 80 adhesive film is recommended and full details of use are provided on
the separate SA 80 data sheet. This adhesive film is supplied on a lightweight glass carrier, or in some cases it can be supplied directly
coated onto one face of the SE 84LV prepreg.
The system is fully compatible with Ampreg wet layup epoxy systems and therefore all types of cores may be bonded to a first skin by
using a separate ‘wet-bonding’ operation. In this case, the addition of filler powders to the appropriate resin system is required to provide
the correct paste-like consistency.
PDS-SE 84LV-27-0421 2
PRODUCT INFORMATION
AVAILABILITY
PRODUCT DESCRIPTION STATUS CERTIFICATION
SE 84LV prepregs are available in a wide variety of fabric forms and
collimated unidirectional tapes. Unidirectional materials are normally SE 84LV HEC, HMC & IMC UD Prepreg
Valid
DNV-GL
145-600g/m2 29 – 40% Resin Content TAK000015K
supplied on a single release paper and fabrics on a single polythene
SE 84LV HEC UD Prepreg Lloyds Register
film. Please contact Customer Support to discuss specific Valid
150 - 600g/m2 LR2165022ALP
requirements and availability. The product formats listed below also SE 84LV IMC UD Prepreg
Valid
Lloyds Register
300-600g/m2 MATS/4888-1
benefit from 3rd Party Certification.
SE 84LV HMC UD Prepreg Lloyds Register
Valid
300-600g/m2 MATS/4887-1
SE 84LV XC Series Prepreg DNV-GL
Valid
290 – 640 g/m2 35 – 45% Resin Content TAK00015Z
SE 84LV XC Series Prepreg Lloyds Register
Valid
XC150 - XC600 g/m² LR2165015ALP
Lloyds Register
SE 84LV RC200T & RC660T Prepreg Valid
LR2165018ALP
Lloyds Register
SE 84LV XE905 Valid
LR2000028ALP
PREPREG PROPERTIES
RHEOLOGY DATA TRANSPORT & STORAGE
SE 84LV resin viscosity profile conducted at 1°C (1.8°F) per minute. When stored sealed & out of direct sunlight.
PROPERTY VALUE STORAGE TEMP UNIT VALUE
Minimum Viscosity 2.9 Pa.s 28.8 P -18°C 0°F months 24
Temperature at Minimum Viscosity 99°C 210°F +18-22°C +64-72°F weeks 8
All prepreg materials should be stored in a freezer when not in

10000
use to maximise their useable life, since the low temperature
COMPLEX VISCOSITY (Pa.s)
reduces the reaction of resin and catalyst to virtually zero.

1000 However, even at -18°C (0°F), the temperature of most
freezers, some reaction will still occur. In most cases after
100 some years, the material will become unworkable.
When not in use SE 84LV products should be maintained at
10 -18°C (0°F). To avoid contamination on their surfaces, allow
rolls to reach room temperature before unwrapping.
1 HEALTH AND SAFETY

30 50 70 90 110
Please refer to product SDS for up to date information specific
TEMPERATURE (ºC)
to this product.
MINIMUM CURE TIME & TEMPERATURE

Recommended minimum cure is 12 hours at 80°C (176°F) using vacuum bag processing.
PROPERTY OVEN / VAC BAG TEST STANDARD
2
Typical Laminate 8 plies of SE 84 LV 300g/m unidirectional prepreg with 35% resin content -
Typical Ramp Rate 1 – 2°C (2 – 4°F) per minute -
Cure Temperature 80°C (176°F) 120°C (248°F) -
Cure Dwell Time 12 (hours) 1 (hour) -
Cure Pressure -1bar (14.5Psi) -
De-mould Temperature < 60⁰C (140°F) -
Dry Tg1 (DMA) 98⁰C / 208°F 115⁰C / 239°F ISO 6721 (DMA)
*suitable for use in conjunction with hot-in / hot-out rapid component manufacture is possible using appropriate press tooling
CURING LARGE STRUCTURES

Gurit provides detailed processing notes for large structures to be built using SE84LV / SA 80; these notes are available from the
Technical Department on request.
PDS-SE 84LV-27-0421 3
MECHANICAL PROPERTIES
Cured using standard vacuum bag processing techniques and a minimum cure time of 12hrs at 80°C (176°F). Values are representative of
the typical properties to be expected but do not constitute a guaranteed specification.
CURED RESIN PROPERTIES
PROPERTY SYMBOL SE 84LV RESIN CAST TEST STANDARD
Tensile Strength T 82 MPa ISO 527-2
Tensile Modulus ET 3.9 GPa ISO 527-2
Flexural Strength F 123 MPa ISO 178
Flexural Modulus EF 3.5 GPa ISO 178
Compressive Strength C 163 MPa ISO 604
Glass Transition Temperature Tg1 115°C ISO 6721
UNIDIRECTIONAL LAMINATE PROPERTIES
Properties presented are averages of multiple batch data from a variety of fibre suppliers. Customers with specific requirements should
contact Gurit technical support who can recommend appropriate fibres and formats.
PROPERTY SYMBOL UNIT HEC FIBRE* IMC FIBRE* HMC FIBRE* UHMC FIBRE* TEST STANDARD
Typical Fibre Density fibre g/cm 3

1.8 1.79 1.82 1.84 -
Fibre Modulus Efibre GPa 227 - 257 275 - 310 365 - 405 420 - 455 -
Resin Content % % 32 - 37 32 - 37 33 - 37 35 ASTM D 3171 Method II
Fibre Volume Fraction t % 55.0 55.5 54.4 54.3 ASTM D 3171 Method II
0° Tensile Strength** XT MPa 2458 2894 2658 1980 ISO 527-5
0° Tensile Modulus** Et GPa 134 170 222 250 ISO 527-5
0° Compressive Strength** XC MPa 1354 1417 1166 1070 SACMA SRM1-94
0° Compressive Modulus** EC11 GPa 121 153 192 227 SACMA SRM1-94
90° Tensile Strength YT MPa 39.2 33.2 30.1 26.0 ISO 527-5
90° Tensile Modulus ET22 GPa 8.3 8.4 7.1 6.6 ISO 527-5
0° Flexural Strength XF MPa 1448 1406 - - ISO 14125
0° Flexural Modulus EF11 GPa 106 129 - - ISO 14125
0° ILSS XILSS MPa 86.6 88.6 82.3 77.8 ISO 14130
*HEC = High Elongation Carbon, IMC = Intermediate Modulus Carbon, HMC = High Modulus Carbon, UHMC = Ultra-High Modulus Carbon
**Normalised to 60% fibre volume fraction
PDS-SE 84LV-27-0421 4
WOVEN LAMINATE PROPERTIES
Properties presented are averages of multiple batch data, where possible witnessed by a third party surveyor on a standard fibre type.
Customers with specific requirements should contact Gurit technical support who can recommend suitable fibres and formats.
PROPERTY SYMBOL UNIT RC200T RC416T TEST STANDARD
-
Resin Content % 42 40 ASTM D 3171 Method II
-
Cured Ply Thickness mm 0.23 0.43 ASTM D792
-
Fibre Volume Fraction % 47 - 53 50 - 59 ASTM D 3171 Method II
0° Tensile Strength* XT MPa 719 1006 ISO 527-4
0° Tensile Modulus* Et GPa 60.6 59.1 ISO 527-4
90° Tensile Strength* YT MPa 662 858 ISO 527-4
90° Tensile Modulus* ET22 GPa 61.6 58.9 ISO 527-4
0° Compressive Strength* XC MPa 759 649 SACMA SRM1-94
0° Compressive Modulus* Ec GPa 58.3 55.6 SACMA SRM1-94
90° Compressive Strength* YC MPa 731 659 SACMA SRM1-94
90° Compressive Modulus* EC22 GPa 59.0 55.2 SACMA SRM1-94
0° Flexural Strength XF MPa 847 895 ISO 14125
0° Flexural Modulus EF11 GPa 51.2 49.4 ISO 14125
90° Flexural Strength YF MPa 857 892 ISO 14125
90° Flexural Modulus EF22 GPa 51.5 50.6 ISO 14125
ILSS M MPa 74.8 55.8 ISO 14130
*Normalised to 55% fibre volume fraction
MULTIAXIAL LAMINATE PROPERTIES
Properties presented are multiple batch data, where possible witnessed by a third party surveyor on a standard fibre type. Customers with
specific requirements should contact Gurit technical support who can recommend suitable fibres and formats.
PROPERTY SYMBOL UNIT XC411 TEST STANDARD
-
Resin Content % 40 ASTM D 3171 Method II
-
Cured Ply Thickness mm 0.43 ASTM D792
-
Fibre Volume Fraction % 47 - 59 ASTM D 3171 Method II
+45° Tensile Strength* XT MPa 1124 ISO 527-4
+45° Tensile Modulus* Et GPa 63.8 ISO 527-4
-45° Tensile Strength* YT MPa 1237 ISO 527-4
-45° Tensile Modulus* ET22 GPa 64.5 ISO 527-4
+45° Compressive Strength* XC MPa 595 SACMA SRM1-94
+45° Compressive Modulus* Ec GPa 62.0 SACMA SRM1-94
-45° Compressive Strength* YC MPa 645 SACMA SRM1-94
-45° Compressive Modulus* EC22 GPa 60.2 SACMA SRM1-94
+45° Flexural Strength XF MPa 815 ISO 14125
+45° Flexural Modulus EF11 GPa 41.5 ISO 14125
-45° Flexural Strength YF MPa 1004 ISO 14125
-45° Flexural Modulus EF22 GPa 57.0 ISO 14125
ILSS M MPa 49.7 ISO 14130
*Normalised to 55% fibre volume fraction

PDS-SE 84LV-27-0421 5
NOTICE
All advice, instruction or recommendation is given in good faith but the selling Gurit entity (the
Company) only warrants that advice in writing is given with reasonable skill and care. No further
duty or responsibility is accepted by the Company. All advice is given subject to the terms and
conditions of sale (the Conditions) which are available on request from the Company or may be
viewed at Gurit’s Website: www.gurit.com/terms-and-conditions.aspx
The Company strongly recommends that Customers make test panels in the final process
conditions and conduct appropriate testing of any goods or materials supplied by the Company prior
to final use to ensure that they are suitable for the Customer’s planned application. Such testing
should include testing under conditions as close as possible to those to which the final component
may be subjected. The Company specifically excludes any warranty of fitness for purpose of the
goods other than as set out in writing by the Company. Due to the varied nature of end-use
applications, the Company does, in particular, not warrant that the test panels in the final process
conditions and/or the final component pass any fire standards.
The Company reserves the right to change specifications and prices without notice and Customers
should satisfy themselves that information relied on by the Customer is that which is currently
published by the Company on its website. Any queries may be addressed to the Technical Services
Department.
Gurit is continuously reviewing and updating literature. Please ensure that you have the current
version by contacting your sales contact and quoting the revision number in the bottom left-hand
corner of this page.
TECHNICAL CONTACT INFORMATION

For all other enquiries such as technical queries:
Telephone + 44 1983 828000 (08:30 – 17:00 GMT)

Email technical.support@gurit.com
24-HOUR CHEMICAL EMERGENCY NUMBER

For advice on chemical emergencies, spillages, fires or exposures:
Europe +44 1273 289451

Americas +1 646 844 7309
APAC +65 3158 1412
E customer.support@gurit.com
W www.gurit.com
All trademarks used or mentioned in this document are protected by law.
PDS-SE 84LV-27-0421 6
XC305
Fabric Description
Gurit Style Unit Width (mm) Style Nominal g/sq.m
XC305 sq.m 1270 +/- 45 Biaxial 302
Fabric Construction
Stitch Length Wt Weight per axis g/m2 Tex per axis

Type
no/cm mm g/m2
0 90 +45 -45 0 90 +45 -45
1.96 2.02 linear 6 1 2 145 145 34 68 400 400
The policy of Gurit is one of continual development and improvement. Gurit reserves the right to alter specifications and prices without prior notice.
Any information or advice obtained from Gurit by other means and whether relating to Gurit material or other materials, is given in good faith.
However, it remains the responsibility of the customer to ensure that Gurit materials are suitable for the particular purpose intended.
EPOXY LAMINATING RESIN
easycomposites
EL2
TM
share the knowledge
Key Features
• Easy to use
• Medium Viscosity Resin
• Outstanding Wetting Abilities
• Good Mechanical Properties
• Choice of Hardener Speed
Product Description Hardener at a ratio of 100 parts of resin to 30 part sof hardener, by weight. FAST
and SLOW hardeners can be blended to achieve pot-life and demould times any-
where between those stated. However, you must still maintain the correct overall
El2 is a high performance general purpose epoxy laminating resin for use in wet- ratio of resin to hardener to ensure a proper cure.
lay or vacuum bagging carbon fibre, aramid or glass lamination.
When working with any epoxy resin, it is essential to mix the resin and hardener
This medium viscosity epoxy exhibits excellent wetting characteristics especially exactly at the correct mix ratio. Failure to do so will result in a poor or only partial
when used with carbon fibre and aramid fibre (such as Kevlar®) reinforcement cure of the resin, greatly reduced mechanical properties and possibly other
making airbubbles in the laminate less likely than with some alternative epoxy adverse effects. Under no circumstances add ‘extra hardener’ in an attempt to
systems. speed up the cure time; epoxies do not work in this way.
The resin also exhibits excellent cured mechanical properies far in excess of many
more traditional epoxy resin brands (as can be seen from the technical data sheets).
Improved mechanical properties mean stronger, ligher, higher performance parts.
Mixing Instructions
EL2 is a highly reactive (fast curing) resin system. Only weigh out and mix as much
resin as you can use within the pot life.
Recommended Uses Weigh or measure the exact correct ratio of resin and hardener into a straight
sided container. Using a suitable mixing stick begin to mix the resin and hardener
Use as a general purpose laminating epoxy, wet-laying or vacuum bagging com- together to combine them completely.
posites such as glass fibre, carbon fibre and aramid fibre (Kevlar). Spend at least one minute mixing the resin and hardener together, paying
particular attention to the sides and base of the container. Remember: Any resin
When cured the epoxy exhibits good flexural strength making it will suited to the that has not been thoroughly combined with hardener will not cure.
lamination of structural parts. The resin also exhibits very good clarity making Once you have finished mixing in one container, it is good practice to transfer the
it also suitable for use when laminating unpainted carbon fibre composites. mixed resin into a second container and undertake further mixing of the resin
Properties
using a new mixing stick. Doing so will eliminate the risk of accidentally using
unmixed resin from the bottom or sides of the container.
The table below shows the typical uncured properties: Pot-Life / Working Time / Cure Time
Property Units Resin Hardener Combined EL2 is a highly reactive resin system and once the resin has been mixed with the
hardener, the reaction will start to give off heat (exotherm) which will further
Material - Epoxy Resin Formulated Epoxy accelerate the cure of the resin, especially when the resin is in the mixing pot.
Amine
Transfer the resin from the mixing pot onto the part as soon as possible to extend
Appearance - Clear Liquid Amber Liquid Clear Liquid
the working time and avoid the risk of uncontrollable rapid cure in the mixing pot.
Viscosity @20 °C mPa.s. 1200 – 1800 5 - 80 1000 – 1400
As with all epoxies, the pot-life/working time will vary significantly depending on
Density @20 °C g/cm³ 1.13 – 1.17 0.90 – 1.06 1.05 – 1.15 the ambient temperature, the starting temperature of the resin and hardener and
the amount of resin mixed.
How to Use
EL2 can be used in ambient temperatures between 15°C (59°F) and 30°C (86°F).
For best results, an ambient temperature of at least 20°C (68°F) is recommended.
Ensure that both resin and hardener containers are within this temperature range
EL2 is a chemical product for professional use. It is essential to read and before use.
understand the safety and technical information before use. The table below gives an indication of pot-life and cure properties:
Follow the guidelines for safe use outlined in the SDS which include the use
ofappropriate hand and eye protection during mixing and use. Pot-Life @ 25 °C Gelation @ 25 °C Demould Time @ 25 °C
AT30 SLOW* 95 - 115mins 8.5 - 10.5hrs 20 - 30hrs
Mix Ratio AT30 FAST* 12 - 17mins 2 - 3hrs 4 - 6hrs
*Fast and slow hardeners can be blended to achieve pot-life and demould
Mix Ratio 100:30 by Weight timesanywhere between those stated above.
EL2 Epoxy Laminating Resin should be mixed with AT30 FAST or AT30 SLOW
EL2 EPOXY LAMINATING RESIN - Technical Datasheet Page 1 of 2

Easy Composites Ltd | Tel +44 (0)1782 454499 | www.easycomposites.co.uk | sales@easycomposites.co.uk
EL2 EPOXY LAMINATING RESIN
Full Cure / Post-Cure Mechanical Properties

As with most epoxy systems, where parts cure in normal ambient temperatures,
full cure is not reached for several days. Although parts will be handleable after Cured Resin Properties
the listed demould time (at 25°C), full mechanical properties will take at least 14
days to develop in (at 25°C). Where possible, avoid exposing the cured resin to full Units AT30 SLOW AT30 FAST
service rigours for at least this time.
Hardness Shore D 84 - 88 85 - 89
As with many post-cure cycles for resins, the post-cure cycle for our EL2 Epoxy
Linear Shrinkage % 0.5 0.5
Resin is not too sensitive and a range of different post-cure cycles will produce
good results, specifically improved mechanical performance and elevated Tensile Strength MPa 70.0 - 80.0 67.0 - 75.0
HDT/operating temperature. Post-curing parts that will be used at or exposed Elongation at Break % 6.0 - 10.0 6.0 - 8.0
to elevated operating temperatures (such as vehicle bonnets/hoods in direct
sunlight, engine-bay parts, car interior parts etc.) is strongly recommended to Flexural Strength MPa 103 - 117 120 - 130
prevent distortion of the parts when they are put into service and experience these Flexural Modulus MPa 2600 - 3200 3600 - 4000
higher temperatures. H.D.T °C 82 - 88 70 - 76
Where possible, parts should be post-cured still inside the mould to reduce
Transport and Storage

distortion and improve surface finish (i.e. reduce ‘print-through’). When post-
curing parts in the mould, it is important to post-cure them without demoulding
at all (i.e. don’t demould and then put them back into the mould) otherwise you
can get some strange patterns on the surface where some areas are post cured in Resin and hardener should be kept in tightly seal containers during transport and
direct contact with the mould surface and others are not. storage. Both the resin and hardener should be stored in ambient conditions of
A simple and very effective post-cure cycle with the EL2 Epoxy Laminating Resin between 10°C (50°F) and 25°C (77°F).
is as follows: When stored correctly, the resin and hardener will have a shelf-life of 12 months.
Although it may be possible to use the resin after a longer period, a deterioration
in the performance of the resin will occur, especially in relation to clarity and cure
CYCLE #1 SUITABLE FOR MOST SITUATIONS profile.
Pay particular attention to ensuring that containers are kept tightly sealed. Epoxy
• 24hrs at room temperature
hardeners especially will deteriorate quickly when exposed to air.
• 6hrs at 60°C
If you’re encountering any surface finish issues (faint print-through) then you can
experiment with a slower ‘ramp rate’ which sometimes improves things:
Disclaimer
This data is not to be used for specifications. Values listed are for typical properties
and should not be considered minimum or maximum.
CYCLE #2 SUGGESTED FOR SUBTLE IMPROVEMENTS TO Our technical advice, whether verbal or in writing, is given in good faith but Easy
SURFACE FINISH Composites Ltd gives no warranty; express or implied, and all products are sold
upon condition that purchasers will make their own tests to determine the quality
and suitability of the product for their particular application and circumstances.
• 24hrs at room temperature Easy Composites Ltd shall be in no way responsible for the proper use and service
• 2hrs at 40°C of the product, nor for the safeguarding of personnel or property, all of which
• 2hrs at 50°C is the duty of the user. Any information or suggestions are without warranty of
• 5hrs at 60°C any kind and purchasers are solely responsible for any loss arising from the use
of such information or suggestions. No information or suggestions given by us
If you need to push the HDT of the finished part higher then you could increase shall be deemed to be a recommendation to use any product in conflict with any
post-cure up to a maximum of 80°C as follows: existing patent rights. Before using any of our products, users should familiarise
themselves with the relevant technical and safety datasheets provided by Easy
Composites Ltd.
CYCLE #3 SUGGESTED FOR HIGHEST POSSIBLE HDT/OP-
ERATING TEMPERATURE
• 24hrs at room temperature

• 2hrs at 40°C
• 2hrs at 50°C
• 2hrs at 60°C
• 2hrs at 70°C
• 4hrs at 80°C
These are all just suggestions. Most situations just call for option #1; 6hrs at 60°C.
Many customers also find that they can dispense with the 24hrs cure at ambient
easycomposites
and simply load newly infused parts into the oven to begin the cure however this TM
is something that you would need to experiment with yourself. A cure at ambient
temperature before post-cure is generally favoured with most resin systems. share the knowledge
Easy Composites Ltd

Unit 39, Park Hall Business Village, Longton, Stoke on Trent, Staffordshire, ST3 5XA, United Kingdom.
Tel. +44 (0)1782 454499, Fax. +44 (0)1782 596868, Email sales@easycomposites.co.uk, Web www.easycomposites.co.uk
EL2 EPOXY LAMINATING RESIN - Technical Datasheet Page 2 of 2

Design and Analysis of Formula Student Race Car Wheel Rims

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Design and Analysis of Formula Student Race Car Wheel Rims

Uploaded by

Copyright:

Available Formats

University Degree in Mechanical Engineering

“DESIGN AND ANALYSIS OF

Pablo Aparicio Alonso

Francisco Javier Velasco López

This work is licensed under Creative Commons Attribution – Non Commercial –

Figure 6.1 Four-spoke rim variant. . . . . . . . . . . . . . . . . . . . . . . . 50

Table 5.1 Mechanical properties for CMCFRP, as per samples tested. . . . . . 46

CAD : Computer Aided Design

CFRP : Carbon Fiber Reinforced Polymer

CMCFRP : Compression Molded Carbon Fiber Reinforced Polymer

FEA/FEM : Finite Element Analysis/Finite Element Method

FS /FS AE : Formula Student/Formula Society of Automotive Engineers

FS G : Formula Student Germany

KEM : Key Efficiency Metric

LT S : Lap Time Simulation

MADFT : MAD Formula Team

PMD : Project Management Document

Prepreg : Pre-Impregnated Carbon Fiber

• Characterize the mechanical properties of short-strand Compression Moulded Car-

• Adapt the design to facilitate low-budget manufacturing within the capabilities of

1.4. Tools and Software used.

• Finite Element Analysis Pre-processing: Altair Hypermesh.

• Finite Element Analysis Solver: Altair Optistruct.

• Finite Element Analysis Post-processing: Altair Hyperview.

• Computer Aided Design: Dassault Systemes SolidWorks.

• Product Lifecycle Management/CAD repository: StrataSYS GrabCAD.

• Mechanical properties analysis, data analysis, graph-making, budget control: Mi-

• Laptime Simulation: OptimumG OptimumLap.

• Project management, Key Efficiency Metric: MADFT Project Master Document.

2.1. Brief history of Formula Student competitions.

Formula Student is a university-level engineering competition in which teams of undergraduate-

2.3. Formula Student Events and Scoring.

• Pre-scrutineering: Checks to the driver safety gear (restraints, fireproof underwear

• Scrutineering (Mechanical): Fulfillment of general rules, safety and structural in-

2.4. MAD Formula Team

Fig. 2.15. Formula UC3M C3R19 in Hockenheim, Baden-Württemberg, Formula Student

3.1. Commercially-developed rims.

3.2. Student-developed rims as state of the art.

Fig. 3.4. UAS Graz rims, courtesy of FSG Media Team.

Older designs by FS teams such as KIT Racing (Karlsruhe Institute of Technology,

JMS Racing (University of Kansas, USA) [Walther, 2016]developed a simple rim

Fig. 3.6. University of Kansas rims.

Fig. 3.7. TU Delft rims, courtesy of FSG Media Team.

4.1. Basics of FS Dynamic disciplines.

As discussed on section 2.2, Formula Student dynamic disciplines constitute 67.5% of

4.3. Lap Time Simulation. Introduction to the Key Efficiency Metric.

4.4. Project Budget: Impact and viability analysis.

LAP TIME [s] SCORE [pts] POINTS DELTA TO BEST [pts]

Table 4.1. BEST TIMES, CORRESPONDING SCORE AND POINTS

Fig. 4.8. Empty KEM tool, before data entry.

5.1. Brief introduction to the Finite Element Method and software.

• Post-processor: Necessary to obtain a visual representation of the results, which

5.1.1. Mesh quality

Finally, connectivity of the elements depends on mesh structure. Well organized

5.2. Loads and load cases.

5.2.1. Maximum Braking.

W = Longitudinal load on each front corner, in N.

5.2.2. Maximum Cornering.

∆W = Lateral load transfer in N

W = Lateral load on each rear corner, in N.

Resulting on a total load of 2362.45 N on the most critical cornering case.

5.2.3. Tyre Mounting.

5.3.1. Dynamic load application.