Chapter 4

Vehicle Structures and Materials

Abstract This chapter commences with a review of chassis structures for the dif-
ferent classes of road vehicles including mass-produced passenger cars, high per-
formance vehicles, small sports cars and commercial vehicles. It proceeds to
consider the different materials used in vehicle structures with a focus on reducing
vehicle weight and therefore emissions through the use of high strength steel, alu-
minium and composite materials. The following section outlines different methods
of analysis of vehicle structures including both traditional theoretical methods and
modern computational techniques. Crash safety of vehicles under impact loading
is then considered and a particular case study of the crashworthiness of a small
space frame sports car presented in detail. The final section of the chapter deals
with the durability assessment of vehicle structures and again includes a detailed
case study of the fatigue assessment and optimisation of a suspension component.

4.1 Review of Vehicle Structures

The purpose of any road vehicle structure is to support all the major components
and sub-assemblies making up the complete vehicle (i.e. engine, transmission,
suspension, etc.) and also carry the passengers and/or payload in a safe and com-
fortable manner.
In the early years of motor vehicles, both passenger cars and commercial
vehicles were manufactured in the traditional way with a separate chassis frame
onto which a non-structural body shell was attached. This form of construction has
survived in commercial vehicles and also in specialist car brands such as Morgan.
Since the chassis frame carries all the applied loads (i.e. dead-weight loads due to
self-weight of vehicle and payload as well as “live” loads due to aerodynamic and
dynamic tyre loads), it must be sufficiently strong and rigid. Most chassis frames
are
of ladder form, that is two longitudinal members connected by a number of
transverse or cross-members which may not all be perpendicular to the
longitudinal members but may take a diagonal or cruciform shape. The body shell
serves mainly as a protection from the elements; it is generally isolated from
the chassis via

© Springer International Publishing AG 2018

215
D. C. Barton and J. D. Fieldhouse, Automotive Chassis Engineering,
https://doi.org/10.1007/978-3-319-72437-9_4
21 4 Vehicle Structures and Materials
6
flexible mountings (usually rubber) and therefore contributes very little to the
overall vehicle stiffness.
The principal drawback of the separate chassis frame is that it is essentially a
2D structure and the members therefore have to be of high section modulus and
are relatively heavy. Moreover it invariably leads to mounting problems due to the
large difference in stiffness between the frame and the body shell. However, this
type of construction is still a popular form of construction in commercial vehicles
since a variety of different forms of body shell can be often mounted on a common
ladder-type frame and the weight of the vehicle structure is of less concern than its
overall load-carrying capability, see Fig. 4.1 for a typical example.
Following World War II, the drive to produce a more efficient structure for
passenger cars led to the development of semi-integral forms of construction.
These retain a strong chassis structure but, by mounting the body shell in a more
rigid manner, the latter is required to carry a proportion of the applied loads. Of
course, the body shell then has to be designed to withstand these loads and even
more attention has to be paid to the mounting points between the frame and body
because of the significant force transmission between the two structures.
Moreover, semi-integral vehicles remain relatively heavy and have to be carefully
assembled to tight tolerances since small misalignments can greatly increase stress
concen- trations near mounting points.

Fig. 4.1 Typical small truck chassis.

Photo Christopher Ziemnowicz, licensed under [CC BY-SA 2.5], via Wikimedia Commons from
“Wikimedia Commons”
 4.1 Review of Vehicle 2
The final shape in the development of the structure of mass-produced passenger
cars was the emergence of unitary or integral forms of construction. As the name
suggests, such vehicles have no discernible separate chassis and the whole body is
designed as an integral unit capable of reacting the applied loads and providing the
necessary stiffness to the vehicle. This form of construction produces a structure
which is truly 3D in the way it deforms and carries load. Thus it can be designed
to be significantly lighter than traditional chassis-framed vehicles because of the
rel- atively large depth of the fabricated structures used to resist the bending and
tor- sional loads. However, since these structures are conventionally manufactured
from relatively thin sheet steel or aluminium, they very often have to be stiffened
with reinforcements or made as box sections. This requires complex tooling and
assembly techniques, the cost of which can only be justified for mass produced
vehicles. Figure 4.2 shows a modern aluminium integral body shell on a robotic
assembly line. The complexity of the body shell as well as the robotic manufac-
turing processes employed in its construction are obvious.
Alongside the mass-produced integral steel and (more recently) aluminium
structured vehicles, a number of alternative forms of construction exist for more
specialised, low volume passenger vehicles. These range from the conventional
chassis plus separate body used on traditional sports cars such as the Morgan, to
the high-tech aluminium plus composite construction employed in the latest high
performance sports vehicles such as manufactured by Lotus. Composites
consisting

Fig. 4.2 Aluminium integral body shell during manufacture.

Reproduced with kind permission of Jaguar Land Rover © Jaguar Land Rover Ltd
21 4 Vehicle Structures and Materials
8
of a thermo-setting epoxy or polyester resin reinforced with glass fibre mat have of
course for a long time been used in the construction of body parts for niche sports
cars (or as replacements for corroded steel parts in mass-produced vehicles).
Recently, more advanced composites using a tougher thermoplastic matrix and
stronger, stiffer fibres such as carbon or Kevlar have been introduced as load-
bearing members in sports cars and other specialised vehicles. Of course, the body
structure of F1 Grand Prix cars and similar relies almost exclusively on the use of
carbon fibre composite materials to give the exceptional stiffness, strength and
lightness of these vehicles.
An important class of vehicle in the small sports car category utilises the
so-called space-frame chassis form of construction. A true space-frame is a col-
lection of tubular members usually welded together and triangulated in such a way
that the members carry load in tension and compression only and do not suffer
bending or torsional to any significant degree. A good example of this form of
construction is the Caterham Super Seven (formerly the Lotus Seven) illustrated in
Fig. 4.3. The main members are predominantly square Sect. 16 swg or 18 swg
steel tubes (with some round tube at certain locations) which are MIG welded
together to produce a strong and stiff frame to which all other components of the
vehicle (engine, suspension, body panels, etc.) are mounted directly. Such space-
frame chassis are widely employed in the kit car market as well as in the
construction of low cost race cars such as in Formula Student/SAE due to the low
cost of tooling for such space frames.
Considering briefly commercial vehicle structures, most of these employ a
separate chassis frame generally manufactured from rolled steel sections of
channel or I-beam form. Two main longitudinal members usually extend the full
length of

Fig. 4.3 Typical space-frame chassis for small sports car.

Reproduced with kind permission of Caterham Cars © Caterham Cars Ltd
 4.1 Review of Vehicle 2
the vehicle. These are joined together by cross members to form a ladder and/or
cruciform shape as illustrated in Fig. 4.1. The power unit, driving controls, axles
and the cab unit are mounted directly on the chassis frame and often form a self-
contained unit. A separate loading platform or a superstructure is also mounted on
the frame according to the usage for which the unit is intended. A strong structural
panel (or bulkhead) normally extends upwards to the full height of the cab to give
protection to the driver and to prevent the load from moving forward under heavy
loading. Articulated vehicles which consist of a tractor unit and a separate trailer
employ a similar form of construction except that only the tractor unit needs to
accommodate the cab and powertrain. Finally, bus structures usually employ a
sturdy 3D frame which is mounted rigidly onto the chassis and to which the body
panels are attached. Roll-over and side impact protection are obviously of great
importance in bus and coach structural design whereas, for heavy good vehicles,
rear and side guards to prevent other vehicles from “submarining” beneath the
chassis structure during crash scenarios are mandatory.

4.2 Materials for Light Weight Car Body Structures

The mandatory requirements for vehicle manufacturers to reduce their fleet

average emissions of CO2 have led not only to developments in powertrains such
as down-sized IC engines and hybrid/electric drives but also to a strong motivation
to reduce vehicle weights. Reduced vehicle mass reduces fuel consumption and
gives enhanced acceleration/deceleration/handling performance. It can also reduce
dam- age to road surfaces and generally enhances road safety because the kinetic
energy of vehicles is reduced. Even commercial vehicles can benefit from
lightweight chassis construction because hauliers can then increase the payload for
a specified Gross Vehicle Weight (GVW).
In terms of integral forms of construction, interstitial-free steels (alloyed with
small amounts of titanium to remove carbon from interstitial sites) have been
introduced to reduce the thickness, and hence weight, of body panels. High
strength low alloy (HSLA) steels (alloyed with manganese to increase strength and
with niobium/vanadium to provide grain refinement) have also been widely used.
More recently advanced high strength steels (AHSS) have been introduced
because of their enhanced mechanical properties including good impact energy
absorbing behaviour (except for the very high strength steels which are
relatively brittle).
These AHSS’s typically include phases of hard steel such as martensite, bainite or
austenite in addition to the more usual ferrite/pearlite microstructure. The resulting
enhanced yield stress enables panel thicknesses to be further reduced but also
causes manufacturing difficulties due to the reduced ductility and the high amounts
of elastic energy stored that can exacerbate problems such as “springback” fol-
lowing forming operations (springback is the tendency of a pressed panel to return
to its original shape after the forming operation).
22 4 Vehicle Structures and Materials
0
One way to mitigate against these forming difficulties is the use of tailor-
welded blanks (TWBs). These consist of multiple sheets of different shapes and
thicknesses that are welded together and then pressed to the required shape. This
means that the thickness can be enhanced where necessary for structural stiffness
but elsewhere can be made thinner to improve formability and further reduce
weight. Optimisation
of the benefits of TWB’s requires detailed understanding and analysis of body
shell in-service loading. Knowledge of the relevant formability properties of the
mate-
rials as well as their costs is also required.
In recent years, much attention has been paid at the high end of the passenger
car market to the replacement of steel body shells with aluminium alloy which has
a much lower density (about 2700 kg/m 3 compared with 7800 kg/m3 for steel). As
with steel, there are many different grades of aluminium alloy and different heat
treatments are possible to give a wide range of material properties with yield
strengths varying from about 70 MPa up to 700 MPa after age hardening. As with
any metallic material, higher yield strength alloys have lower ductility and are
therefore more difficult to form. For this reason, advanced body shells often
employ a mixture of pressed sheet, extruded and cast aluminium components,
such as in the Audi design shown in Fig. 4.4 which utilises steel, carbon fibre
composite and even magnesium components as well as aluminium to form a
hybrid structure which can be regarded as a space-frame. Both aluminium and
magnesium alloys can be

Fig. 4.4 Audi A8 multi-material body shell.

Reproduced with kind permission of Audi © Audi AG
4.2 Materials for Light Weight Car Body Structures 221

pressure die cast (with or without vacuum assistance) which produces intricate
parts with good properties especially in relation to dynamic loading.
Aluminium alloys can be laser, spot or metal inert gas (MIG) welded and are
also very suitable for adhesive bonding. Higher strength alloys (e.g. the 6000
series commonly used in aerospace applications) are being increasingly used,
often with quick hardening compositions so that age hardening can occur during
the coating or painting processes. Developments in the properties of cast alloys
have led to the consideration of aluminium for demanding applications such as the
frontal impact longitudinal rails shown in blue in Fig. 4.4 and even in brake rotors
where heavy ferrous materials are currently used.
Apart from its higher cost in comparison to steel, a further issue with
aluminium is that it is more difficult to recycle. Again special alloys and processes
have been developed to facilitate recycling e.g. the RC5754 alloy used extensively
by Jaguar Land Rover in body shells such as shown in Fig. 4.2 contains up to 50%
recycled material. In fact Jaguar aims to be using 75% recycled material in its
aluminium body shells in the near future.
Alongside the developments in high performance metallic materials, polymeric
matrix composites are being increasingly used in chassis structures. Over the past
10 years or so, the low cost, low tech glass fibre and the more expensive, high tech
carbon fibre technologies have converged so that high performance composites
have become an affordable option for body parts on normal road cars. There is
now a wide choice of reinforcement types (short or continuous, randomly or fully
aligned carbon, glass or aramid fibres) in both thermosetting (epoxy, polyester,
vinyl ester) and thermoplastic (PP, PEEK) polymer matrices. Composite
production processes have become more automated with much shorter cycle times.
Prominent amongst these processes is Resin Transfer Moulding (RTM) whereby
fibre rein- forcement preforms are infiltrated by liquid polymer under carefully
controlled conditions to achieve the near final net shape within a short cycle
time.
Since complete composite body shells are likely to remain the reserve of high
performance sports and race cars, an issue for more mass produced vehicles is the
joining of a composite component to its metallic neighbours within a hybrid body
shell. Since composites generally cannot be welded, the obvious joining technique
is adhesive bonding but such joints need to be carefully designed and tested to
demonstrate their integrity. An alternative is to incorporate metal fixing devices
within the composite during the manufacturing process.
The modern vehicle structure designer is faced with a wide choice of different
materials and manufacturing techniques. Not only must the complete body shell be
lightweight and sufficiently stiff and strong, it should also be cheap and relatively
easy to manufacture. It most also aim to protect the vehicle occupants and other
road users from serious injury during road traffic accidents. From an
environmental point of view, it is important to reduce the vehicle weight whilst at
the same time improving the carbon footprint and recyclability of the materials
used in its con- struction. Advanced computer simulation techniques exist which
enable designers to predict both the manufacturability and on-road performance of
the materials and
 2 4 Vehicle Structures and

components selected. Overall, with this seemingly never-ending choice of

materials and manufacturing techniques, it is an exciting and challenging time for
the vehicle structure designer.

4.3 Analysis of Car Body Structures

4.3.1 Structural Requirements

The structural requirements of any vehicle structure can be summarised as follows:

1. The structure must be sufficiently stiff to react the static loads (i.e. mainly due
to dead-weight) and dynamic loads (i.e. mainly due to driving over rough
terrain and handling manoeuvres) without excessive deformation.
2. The structure must be sufficiently strong to withstand many cycles of the
applied loading without suffering from fatigue or other forms of material
failure.
3. The structure should deform in such a manner under impact load conditions so
as to minimise the risk of injury to the occupants and other road users.
The first job of any vehicle structure designer is to ensure that requirement 1 is
met since this is critically dependant on the overall structural layout and
determines how the vehicle will handle during normal and extreme manoeuvres.
Basically, in order for the suspension system to perform the tasks it is designed
for, its mounting points should remain as stationary with reference to the vehicle
axis system as possible. There are three main categories of dynamic loading which
tend to deform the vehicle structure about these mounting points:
i. Torsional loads such as when one wheel hits a bump in the road and the
additional suspension load at that corner of the car is transmitted through the
vehicle structure.
ii. Bending loads such as when both wheels on an axle hit a bump or a kerb
simultaneously.
iii. Longitudinal/lateral loads due to inertia effects under traction, braking and/or
cornering conditions, including the in-plane effects of minor collisions and
impacts.
Although in principle all three of these loading types can act simultaneously,
the first of these is the most significant and the latter relatively unimportant in
terms of the chassis structure itself. Thus the torsional stiffness of the structure is
perhaps the most important parameter to be considered in the initial stages of a
chassis design. It is normally measured or calculated by fixing the suspension
mounting points at three corners of the car and applying a vertical load at the
fourth corner as indicated
 4.3 Analysis of Car Body 2

θ
δ

Fig. 4.5 Calculation of chassis torsional stiffness

Fig. 4.6 Computer simulation of space-frame chassis under torsional loading.

Reproduced with kind permission of Simpact Engineering © Simpact Engineering Ltd
 2 4 Vehicle Structures and

Fig. 4.7 Torsional test of space-frame chassis.

Reproduced with kind permission of Simpact Engineering © Simpact Engineering Ltd.

in Fig. 4.5. The applied torque (i.e. P · W ) divided by the angle of twist (i.e.
h ¼ d=W ) is the torsional stiffness Kt as given by Eq. (4.1) below in units of
torque per radian of twist (but normally presented in units of N m/degree).

T P·W P·W2
Kt ¼ ¼ ð4:1Þ
¼ h d=W d

Although a high torsional stiffness is desirable, there is usually a weight

penalty involved in stiffening a vehicle structure. Therefore a suitable parameter to
be optimised at the design stage is the torsional stiffness per unit weight. Also a
very high stiffness may lead to NVH problems with little attenuation of
suspension load variations arising from running on rough roads and problems at
the support points of the more flexible body members.
A computer simulation of a typical space frame under torsional loading is
shown in Fig. 4.6. The same chassis under torsional test in the laboratory is
shown in Fig. 4.7. Note how the chassis is loaded through the front suspension
members but that the shock absorbers have been replaced by rigid struts. The
unique feature of this particular sports car space-frame is that it uses double-butted
welded joints throughout which enables the tube thickness to be reduced where
appropriate and
 4.3 Analysis of Car Body 2
the overall chassis weight to be minimised, albeit at greater material and manu-
facturing cost.
Generally speaking, if a car structure is sufficiently stiff in torsion, it will be
adequate in bending. Also, stresses in individual structural members during
normal driving conditions will be low. However, requirement 2 in the list above
(i.e. to maximise the fatigue life) remains a priority for the designer of the
structure since stress concentrations can lead to material cracking or even failure,
particularly at spot-welded or bolted connections. This fatigue mode of failure is
considered in Sect. 4.5 below. The third and final requirement (i.e. to ensure the
vehicle is crashworthy) often causes conflicts because a strong, stiff structure may
not possess the necessary energy absorbing capabilities. Vehicle crashworthiness
is the subject of Sect. 4.4 of this chapter.

4.3.2 Methods of Analysis

4.3.2.1 Simple Bending Analysis

Whether the purpose of the analysis is to determine the torsional stiffness or the
stresses in the structure due to the applied loads, it is important to establish the
load path through the structure, i.e. how the loads are input to the structure and
transmitted from one member to the next. For example, for the analysis of static
vertical loading, the distribution of loads and centres of mass of major items, e.g.
engine, passengers, payload, etc., need to be determined or estimated. Usually the
weight of the body shell, that is the sprung weight, is assumed to be uniformly
distributed over the length of the vehicle. Then, assuming the vehicle to be simply
supported at the wheels, the type of load diagram shown in Fig. 4.8 can be
constructed leading to derivation of the shear force (SF) and bending moment
(BM) diagrams as shown. It can be seen from these diagrams that the maximum
SF occurs at A and the maximum BM at B.
Note that the static loads are often multiplied by factors of 2 or 3 to allow for
dynamic effects as the vehicle traverses rough ground or hits obstacles in the road.
However this is very much a “rule of thumb” and more precise methods of esti-
mating the effect of dynamic loading are now available (see Sect. 4.5 below).
Once the maximum bending moment has been estimated, the maximum
bending stress in the case of a chassis framed vehicle can then readily be
calculated from the standard equation for a simple beam in bending:

M·y
r¼ I ð4:2Þ

where
r is the maximum bending stress, which should be less than the design stress,
M is the applied bending moment,
 2 4 Vehicle Structures and

Front Engine Rear EngineFront Rear

MountingMountingPassengersPassengers

Wheel Base

Overall Length

Fig. 4.8 Indicative static load diagram for saloon car: “A” indicates maximum shear force (S.F.),
“B” indicates maximum bending moment (B.M.)
 4.3 Analysis of Car Body 2
I is the second moment of area of the frame longitudinal member at that point y is
the maximum distance from the neutral axis of the longitudinal member to its
upper or lower surface.
For a vehicle of integral construction, the calculation is more complex because
many parts of the structure are involved in reacting the applied loads. As a first
approximation, the bending moment can be assumed to be carried solely by the
sills, the longitudinal members usually of folded sheet steel construction which
connect the front and rear of the car underneath the doors. Even with this
assumption, the calculations are not straightforward because the sill section is
often unsymmetrical and therefore subject to warping as well as bending. It should
also be remembered that the torsional stiffness of beam sections with longitudinal
cut-outs to allow for wiring runs e.t.c., is much reduced compared to that of the
corresponding fully closed section. The theory of unsymmetrical sections and
open sections in bending and/or torsion is outside the scope of the present book.

4.3.2.2 Simple Structural Surfaces (SSS) Method

An important concept in the design of integral structures is that of the shear panel.
This is an idealisation of the large, approximately flat, thin steel panels used in the
construction of many vehicles such as vans. It is assumed that a shear panel
transmits load primarily by developing in-plane shear stresses due to shear forces
applied parallel to the edges of the panel. To illustrate how this concept works,
consider the simple box structure subject to a torsional load, T as shown in
Fig. 4.9a. Free body diagrams can be drawn for each panel and the only loads

Q1
Q3

(b)
Q3
Q3 Q1
Q1
(a)
Q2 Q2
Q1
Q2 Q1
L2T Q2
Q3
Q2
L3 Q1
L1 Q3
Q3

Q3
Q1

Fig. 4.9 Simple structural surfaces analysis of simple box structure a full structure b structure
subdivided into panels with edge shear forces indicated
 2 4 Vehicle Structures and

involved are assumed to be the shear forces, Qi acting at the panel edges as shown
in Fig. 4.9b. Then, for static equilibrium of the end, sides and top/bottom panels
respectively, we can write:

Q1L2 þ Q2L1 ¼ T ð4:3Þ

Q3L2 — Q2 L3 ¼ 0 ð4:4Þ

Q3L1 — Q1L3 ¼ 0 ð4:5Þ

From Eq. (4.3):

L3
Q¼ Q ð4:6Þ
3 1
L1
Then, from Eq. (4.2):
L2 L2 L3 L2
Q ¼ Q ¼ · Q ¼ Q ð4:7Þ
2
L3
3 L3 L1 1 L1
1

Finally, from Eq. (4.1):

T L1 T L2
Q ¼ —Q · ¼ — ·
Q L1
1 2 ·
L2 L2 L2 L1 1
L2
Therefore:

T
Q1 ¼
2L ð4:8Þ
2
and Q2, Q3 can then be found from equations Eqs. (4.6 and 4.7).
Note however that, if the top (roof) of the box structure is not present, then
Q1 = Q3 = 0 since there is nothing to react these shear forces on the top of the
structure. Therefore, by the principle of complementary shear, Q2 = 0. The shear
panel concept then falls down and the structure carries the torsional load by a
different mechanism (warping) and will be very much less stiff as a consequence.
This illustrates the difficulties that engineers are faced with when designing an
open-topped convertible car to have sufficient torsional stiffness. Of course, prac-
tical vehicle structures are very much more complex than the above simple
example and the door frame structures (which can be idealised as simple beams)
and even the windscreen can add significantly to the torsional stiffness of a
vehicle.
 4.3 Analysis of Car Body 2
4.3.2.3 Finite Element Analysis (FEA)

It will be appreciated from the above general discussion that the analysis of
vehicle structures using traditional theoretical methods is complex and often very
approximate in view of the number of simplifying assumptions that need to be
made. An alternative approach which has become almost universally adopted in
the industry is to use finite element analysis (FEA) to more accurately predict
defor- mations and stresses in vehicle structures. FEA essentially involves
breaking the
structure down into a large number of non-overlapping small regions (“elements”).
The problem is formulated in terms of the “degrees of freedom” (usually dis-
placements) at certain discrete points known as “nodes” which always occur at
element vertices but may sometimes also be located elsewhere.
There are many different types of elements but the ones most commonly employed
in vehicle body structural analysis are beam and shell elements. Since both these
types only explicitly model the mid-surface of the structure, the section properties
(e.g. second moment of area, shell thickness) are specified separately. A beam ele-
ment model of a space-frame sports car chassis is shown in Fig. 4.10 whereas a basic
shell element model of a pick-up vehicle structure is shown in Fig. 4.11.
Note that in Fig. 4.10, the engine and final drive have been modelled with solid
elements, the properties of which are calculated to give the correct mass distribu-
tion. The chassis members are modelled with beam elements, shown as straight
lines in Fig. 4.10, which have 3 translational and 3 rotational degrees of freedom at
each node. If the connections at the nodes can be considered to be pin-jointed,
then the members carry loads in tension/compression only. However, if the
connections are sufficiently reinforced to be considered rigid, the members can be
subjected to bending/torsion as well as tension/compression. In practice it usual to
make the rigid-joint assumption since this will generally overestimate the stiffness
and the stresses in the members and so lead to conservative (i.e. safe) designs.
The shell element model of the pick-up truck in Fig. 4.11 is actually derived for
a dynamic modal analysis of the body shell and is only indicative of the much
more detailed model that would typically be used for structural assessment. The
front windscreen is included because it does have a stiffening effect on both the
static and

Fig. 4.10 Beam FE model of spaceframe

 2 4 Vehicle Structures and

Fig. 4.11 Shell FE model of pick-up vehicle structure.

Source Advances in Vibration Engineering and Structural Dynamics, Intech 2012.
http://www.intechopen.com/books/advances-in-vibrationengineering-and-structural-dynamics

dynamic behaviour but the side windows and any sun-roof would normally be
excluded from a structural model as they may be open during critical manoeuvres.
Occasionally, full 3D solid elements are used in place of shells e.g. in the detailed
analysis of critical regions such as spot welded connections where stress concen-
trations may lead to early fatigue failures. Normally, isotropic linear elastic material
properties are assumed for steel or aluminium components (only Young’s modulus
and Poisson’s ratio required) but occasionally non-linear plasticity characteristics
will be specified in heavily stressed regions where the material may be subject to
plastic (permanent) deformation. Composite materials are more difficult to analyse
because their properties are often anisotropic (different in different directions of
loading). Modern FEA packages have specialist composite material models avail-
able but it is outside the scope of the present book to describe these in detail.

4.4 Safety Under Impact

4.4.1 Legislation

The safety of vehicles can be divided into primary (sometimes called “active”) and
secondary (“passive”) safety. Primary safety involves the prevention of accidents
and collisions and is the concern of the designers of the suspension and braking
systems in particular e.g. advanced chassis control (ACC) and anti-lock braking
(ABS) are two systems which have been developed to improve primary safety.
Secondary safety is concerned with minimising the risk of injury to the vehicle
occupants and other road users in the event that an accident should occur. As such,
it is very much the responsibility of the vehicle structure designer to ensure that
the
 4.4 Safety Under 2
vehicle has sufficient energy absorbing capability and general “crashworthiness” to
meet the ever increasing legislation and customer-driven demands.
The most significant legislative codes affecting the international design of
vehicles are the Federal Motor Vehicle Safety Standards (FMVSS) which apply to
all vehicles sold in the USA and the various European Council (EC) directives on
vehicle safety which are mandatory for volume-produced vehicles sold in the EU.
Non-regulatory assessments also exist in many jurisdictions under the heading of
NCAPs (New Car Assessment Programmes). Although passing the various NCAP
tests is non-mandatory, obtaining good (ideally 5 star) ratings are important per-
formance targets for vehicle manufacturers.
There are many parts within these standards relating to occupant protection and
component design for frontal impact safety. The original and most well-known
requirement is contained in FMVSS 208 which specifies a maximum allowable
deceleration of 60 g measured on anthropomorphic test dummies in frontal
impacts at 30 mph with a rigid wall at angles between 0° and 30° to the forward
direction of motion. This original requirement was criticised as being
unrepresentative of true crash situations and has been supplemented by tests
involving a deformable barrier (manufactured from aluminium honeycomb to
represent the front-end stiffness of a typical saloon car) and at various degrees of
overlap (typically 40 or 60%) with the front-end of the test vehicle. A computer
simulation of such an offset barrier frontal impact test is shown in Fig. 4.12.

Fig. 4.12 Computer simulation of frontal impact test.

Source Wikipedia Commons. https://commons.wikimedia.org/wiki/File:FAE_visualization.jpg
 2 4 Vehicle Structures and

These offset deformable barrier tests also forms the basis of the current EU
stan- dards although the impact speed specified in Europe is 56 km/h. However the
requirement for a maximum deceleration of 60 g for any time period longer than 3 ms
as measured on instrumented dummies has remained the norm. As well as
dummies representing the average male driver, dummy tests are now also carried out
for smaller female and child occupants and in the rear as well as the front of the
vehicle. Close attention is paid to injury modes such as neck whiplash and lower
leg injuries.
In addition to rigorous frontal impact requirements, side impacts are also the
subject of legislation. Here worst case tests tend to replicate the side of the vehicle
hitting a rigid pole rather than the deformable front end of another vehicle. Since
there is very little room for deformable energy absorbing devices in door
structures, stiff side impact beams are important in protecting the occupant space
from excessive intrusion and close attention is paid to collisions between the
dummy (especially the head) and the surrounding interior structures.
Collisions between passenger cars and larger vehicles such as trucks has also
been the subject of much scrutiny and legislation. Trucks are now required to have
rear and side guards to prevent smaller vehicles “submarining” underneath the truck
chassis on impact with obvious dire consequences for the car passengers as men-
tioned in Sect. 4.1 above.
Finally, protection of pedestrians and other road users (e.g. cyclists) in the
event of a collision with a vehicle is now the subject of legislative standards and
con- sumer tests. Typically instrumented leg and head forms representing both
adult and child anatomy are launched at different angles against the bumper or
bonnet structures of the vehicle. Impact speeds are typically 35 km/h and potential
injury is assessed using standard injury criteria such as AIS and HIC. These tests
have required vehicle manufacturers to rethink the front end designs of their
vehicles to make them far less damaging to pedestrian impacts by removing sharp
changes in section and reducing the stiffness of components such as the bonnet
which are likely to be involved in the collision.

4.4.2 Overview of Frontal Impact

In the frontal impact of a vehicle whether with a barrier or another vehicle, the
worst case scenario is that all the initial kinetic energy of the vehicle on impact is
dissipated within the structure of that vehicle alone. If it is assumed that F is the
force to crush (plastically deform) the front end of the vehicle and s is the crush
distance, then a simple energy balance gives:

ZSf 1
F ds ¼ mv02 ð4:9Þ
2
0
 4.4 Safety Under 2
where
m is the mass of the vehicle
v0 is the velocity on impact
sf is the crush distance at the end of the impact (total front end deformation).
Now, if it is further assumed that the crush force is constant and known, then
the total crush distance is given by:

mv20
sf ¼
2F ð4:10Þ

For example, if a vehicle of mass 1000 kg and crush strength 300 kN impacts a
rigid barrier at 50 km/h (13.9 m/s), the total amount of front-end crush is:
1000 × 13:92
sf ¼ ¼ 0:32 m
2 × 300 × 103

The time duration of the impact, t, can be calculated from the momentum
equation:
F · t ¼ mv0 ð4:11Þ

Giving for the above parameters: t ¼ 1000×13:9 ¼ 0:046 s

300×
The average deceleration of the vehicle to the rear of the crush zone can then be
calculated from the change in velocity divided by the impact time duration:

13:9
a ¼ 0:046
,
¼ 302 m s2
≈ 30 g

The above equations clearly illustrate that a very strong front end structure will
give a short crush length and impact time duration and therefore a high
deceleration. In practice the situation for any real vehicle is far more complex
than suggested by the above equations. In particular, a vehicle absorbs energy
by a variety of mechanisms which include crushing, folding, buckling and
frictional contact of a number of discrete components within the vehicle structure.
The front longitudinal chassis rails in particular, along with other support
structures, are designed to absorb energy progressively as the impact load is
transferred to the rear of the
vehicle (see Fig. 4.13).
Thus the crush strength is not a constant value as assumed above and the
deceleration time history (the “crash pulse” as it is known) is complex, as shown by
the representative sample in Fig. 4.14. Incidently this pulse only just meets the
standard criterion of a maximum deceleration at the occupant restraint position of
30 g for any time period greater than 3 ms.
 2 4 Vehicle Structures and

Fig. 4.13 Load path for frontal impact of vehicle.

Reproduced with kind permission of Audi © Audi AG

Fig. 4.14 Representative Engine contact

frontal impact crash pulse

Front end fully collapses

Vehicle brought to rest

Bonnet crumples
Bumper starts to crush

Simple theoretical analysis will not predict such a complex response which
requires either a full scale crash test or advanced numerical simulation of the
impact event to be carried out. The former is extremely expensive and it is not
easy to investigate the complex interactions that occur within a vehicle structure or
the effect of changing certain parameters. The latter numerical approach to
crashwor- thiness assessment prior to final impact testing of a vehicle has become
the norm as described in the case study presented in Sect. 4.4.4 below.
 4.4 Safety Under 2
4.4.3 Energy Absorbing Devices and Crash Protection
Systems

An ideal energy absorbing (EA) for front-end crash protection has the dynamic
force-displacement (P-d) characteristics shown in Fig. 4.15.
Up to the critical load P c, the deformation should be elastic, i.e. it should be
recovered completely on unloading from a relatively low load (low speed impact).
Once Pc is exceeded, the device should deform at constant load, progressively
absorbing the energy of impact and restricting the level of force transmitted to other
parts of the vehicle. The amount of energy absorbed is of course equal to the area
under the force-deflection (P-d) curve.
The front bumper of a passenger car is normally the first component to be
impacted in a frontal collision. Bumper design has evolved from a simple steel
construction designed to protect the bodywork from small knocks towards the
complex polymer/foam/metal construction of modern bumper systems. Polymer
foams, or honeycomb cellular materials, are widely used as impact energy
absorbers in devices such as knee bolsters and side impact protection systems as
well as for bumpers. Although far from behaving in the ideal manner suggested by
Fig. 4.15, polymer foams have many desirable characteristics. They are light,
corrosion resistant, responsive to any direction of loading, recover completely
when subject to moderate loading below the foam yield stress and can be
engineered to meet the particular requirements by varying the density, type and
form of the polymer construction. Typical load-deflection curves for polymer
foams of varying degrees of porosity (% density of the solid polymer) are shown
in Fig. 4.16.
It can be seen from Fig. 4.16 that after the initial elastic response and yield, the
load-deflection response becomes approximately linear again (linear strain hard-
ening). The response eventually stiffens as the pores in the cellular material are
flattened. However, the energy absorption capabilities of the bumper system can
be extended by mounting the bumper on the front longitudinal chassis rails of the
vehicle by means of a controlled deformation structure. One such arrangement is
the concentric interference-fitted “tube-within-tube” device shown in Fig. 4.17
which is amenable to theoretical analysis as outlined below.

Pc

Fig. 4.15 Ideal EA force-deflection response

 2 4 Vehicle Structures and

P 75%

50%

25%

Fig. 4.16 Typical force-deflection plots for cellular polymer foams of different densities
(expressed as percentage of fully dense polymer)

Fig. 4.17 Concentric tube energy absorbing device

The concentric “tube-within-tube” arrangement shown in Fig. 4.17 absorbs

energy both in plastically expanding the outer tube but also in overcoming friction
between the two tubes as they slide relative to one another. Thus, for small relative
movement dx, the energy absorbed, dE, is:
 4.4 Safety Under 2
dE ¼ dWp þ dWF
Z
ð4:12Þ
¼ dV rde þ F · dx

where
dWp is the plastic work done
dWF is the frictional work done
dV is the volume of material undergoing plastic deformation
F is the axial friction force, assumed to remain constant over the small axial
movement dx.
Now dV ¼ p dm t dx
where dm is the mean diameter of the outer tube and t is its thickness.
Assuming perfectly plastic behaviour and that only the outer tube deforms:
Z
rde
o ¼ Y · etr ¼ Y ln d ð4:13Þ
di

where Y is the yield stress (assumed constant), etr is the mean true strain and do
and di are the external and internal diameters respectively of the outer tube.
ðNote: dm ¼ ðdo þ diÞ=2; t ¼ ðdo — diÞ=2Þ

Furthermore:
F ¼ l pA ð4:14Þ

where
l is the coefficient of friction
p is the radial pressure between the two tubes
A is the area of interference fit (which increases with length of the tube overlap).
Therefore:
do
dE ¼ p dm tY ln d dx þ l · p po x dx ð4:15Þ
di

Thus, the total energy absorbed in increasing the length of interference from l1
to l2 is given by:
Zl
do Zl2
2
p tY ln · dx lpp x · dx
E¼
dm þ l1
do ð4:16Þ
l1 di l2 — l2
d 1
¼ p dm tY ln
o
ðl2 — l1Þ þ l p p do 2

di 2
 2 4 Vehicle Structures and

=Y =Y

Fig. 4.18 Section though the outer tube of concentric tube EA device

Finally, p can be related simply to Y assuming the entire wall thickness of the
outer tube is plastically deformed as shown in Fig. 4.18.
For static equilibrium: pdox ¼ 2Y tx
Therefore:
2Yt
p¼ ð4:17Þ
do

Then, if Y and l are known or can be estimated, the rate of energy absorption
can be calculated for any given set of geometric parameters (di, do, l1) and these
parameters optimised to give the desired performance.
A typical experimental load-deflection curve for the concentric tube device is
shown in Fig. 4.19. Although this is not quite the ideal response as shown in
Fig. 4.15, the fact that the device can be reliably tuned, and even re-used if the
impact is not too severe, makes it a potentially useful means of limiting the force
transmitted from the bumper to the main vehicle structure.

Fig. 4.19 Axial load-deflection curve for concentric tube EA device

 4.4 Safety Under 2
4.4.4 Case Study: Crashworthiness of Small Spaceframe
Sports Car

As mentioned above, many small sports and racing cars employ a spaceframe in
the form of a strong and stiff tubular steel chassis. There is very limited scope within
such spaceframes for energy absorption by the large deformation of sheet steel body
panels which occurs in mass-produced integral body structures. Furthermore, for
styling reasons, it is often not considered possible to mount large bumpers or other
forms of external energy absorbing devices at the front of the vehicle. Meeting the
crashwor- thiness legislation is a major challenge for any such vehicle and the
research described below aimed to investigate the crashworthiness of a typical small
spaceframe vehicle using finite element simulation of the impact of the whole
vehicle against a rigid barrier as well as drop weight testing of individual
components.
For most small sports/race cars, the wheels are external to the body and, in a frontal
impact, the front tyres very quickly came into contact with the barrier, unlike in a
normal passenger car where contact between the tyre and front wheel arch occurs
much later during the impact event. The overall response therefore depends very
much on the impact characteristics of the tyres as well as the wheels and
suspension system on which they are mounted. Much effort was devoted in the
case study to modelling the tyre and wheel in as realistic yet relatively simplistic
manner as pos- sible. This involved developing an air-bag model for the inflated
tyre and using a finite element model to simulate the experimental impact of the tyre
and wheel against a rigid block using a drop test carriage to house the assembly as
indicated in Fig. 4.20.

Fig. 4.20 Finite element model of tyre and wheel mounted in drop cage
 2 4 Vehicle Structures and

Force

Time (ms)
Displacement

Experimental results Time (ms)

Simulation results

Fig. 4.21 Comparison of measured and predicted force-and deflection-time response of tyre when
impacted against rigid barrier

The results were validated by comparison with instrumented drop tests on

actual sports car wheels and tyres, as indicated in Fig. 4.21. The force was
measured by means of a load cell within the impact anvil and the deflection of the
outer tyre wall by means of a high speed video camera. The agreement between
test and simulation was considered sufficiently close, especially with respect to
energy absorbed by the tyre, to use the model with confidence for the full vehicle
impact simulations which are described below.
Apart from the tyres and wheels, the investigation concentrated on modelling
the impact deformation of the main spaceframe chassis members with either
beam or
 4.4 Safety Under 2

Deformation
experience during
bending of tubes.

Fig. 4.22 Deformed shape of representative pyramid structure after impact—test piece (left side)
and shell element simulation of one quarter of structure (right side)

shell elements. Again the predicted results were validated by testing the dynamic
response of representative “pyramid” welded steel structures in a large drop test
facility. A comparison between the post-impact experimental and numerical
shapes for these simple pyramid structures is shown in Fig. 4.22. It was found that
to accurately model the plastic hinge formation and subsequent ovalisation of
the
longitudinal steel tubes, it was necessary to have a much finer mesh in the local
area of the hinge than the mesh indicated in Fig. 4.22.
The complete finite element model of a particular space-framed vehicle shown
in Fig. 4.23 makes extensive use of shell-type elements of sufficient mesh density
at the front end of the vehicle. Elsewhere simple beam elements are used and the
model includes 3D representations of the engine and other heavy components in
order to obtain the correct mass distribution. A dynamic analysis using an explicit
finite element solver was carried out for this model at an impact velocity of 50
km/h against a rigid barrier. This gave predictions of deformation as shown in
Fig. 4.24
which agree well with the limited full car crash test results available. The
predicted deceleration time history, monitored at a point equivalent to the driver’s
chest, has the form shown in Fig. 4.25. The peak deceleration of 87 g is well in
excess of the usual limit of 60 g. The incorporation of EA devices such as
described above, either within the vehicle nose cone (light plastic structure not
included in model) or as an
integral part of the chassis structure, was proposed as a potential means of
improving the frontal impact characteristics of the vehicle without affecting its
overall appearance.
 2 4 Vehicle Structures and

Fig. 4.23 FE model of small space-frame sports car—undeformed

Deformation of
tyres during impact

Bending of
chassis members during impact

Fig. 4.24 Deformed shape of FE model of small space-frame sports car during frontal impact
 4.5 Durability 2

Deceleration

Time (ms)

Fig. 4.25 Predicted crash pulse at driver restraint position for frontal impact of small space-frame
sports car

4.5 Durability Assessment

4.5.1 Introduction

As discussed above, automotive structures and components rarely fail due to a

single application of loading, except of course when they are designed to do so
such as in impact crash scenarios. Instead structures and components may
ultimately fail over a prolonged service life due to the accumulation of fatigue
damage coupled with environmental effects such as corrosion and wear which
reduce the load carrying capacity of the material. The study of these long-term
loading effects is
known as “durability assessment” and is considered an important part of the overall
structural reliability of the vehicle.
Experimental testing for durability is carried out in various ways:
• in-service field testing
• accelerated proving ground tests
• laboratory testing.
Field testing with instrumented vehicles is the most realistic form of testing but
it is very time-consuming and expensive. In fact, for a well-designed vehicle, no
failures should be generated in a reasonable timescale unless the vehicle is sub-
jected to extreme misuse load cases such as hitting kerbs or driving over severe
pot-holes. Nevertheless the local acceleration and strain data generated can be
useful in other more detailed forms of assessment.
 2 4 Vehicle Structures and

Fig. 4.26 Four poster dynamic test rig

Proving ground tests normally involve driving the vehicle over special tracks
incorporating pavé (cobblestone-like) or corrugated surfaces. These are designed
to subject the vehicle to more extreme loading than would be the case on normal
roads and hence promote failures in a shorter time scale. Again the vehicle will
usually be instrumented with accelerometers and/or strain gauges at critical
locations to pro- vide useful data on loading conditions.
Finally, laboratory testing is a means of subjecting components and sub-
assemblies to realistic loading under carefully controlled conditions. For example,
4 post dynamic test rigs can be used to load all 4 wheels stations of a vehicle
independently, as indicated in Fig. 4.26. Since the tests can be run con- tinuously
over an extended period, this can be a cost-effective and reasonably rapid means
of establishing the durability of automotive structures and components. However
the capital investment in test equipment is high and prototype parts and assemblies
must be available. This means that testing can only occur quite late in the product
design cycle and cannot therefore be used to inform or drive design changes
leading to improved solutions.

4.5.2 Virtual Proving Ground Approach

In order to bring more reliable and efficient vehicles to the market place more
rapidly, it is highly desirable to substitute or supplement the above traditional
durability test methods with predictive software tools. This has led to the concept
of a Virtual Proving Ground (VPG) in which computer modelling techniques
replace physical tests. Such an approach should enable alternative designs to be
assessed far earlier in the product life cycle than previously leading to more
efficient and novel solutions being generated. As well as being more cost effective,
this could also pave the way for more automated and optimised design methods
which may replace some of the reliance on the accumulated experience of
previous generations of engineers and/or tried-and-tested solutions.
 4.5 Durability 2
The heart of any VPG approach to durability assessment is accurate fatigue
data for the material or component under consideration, usually in the form of the
classic S-N curve where S is the stress (or strain) amplitude and N is the number
of load cycles before failure. The relationship should ideally contain design factors
to allow for the variability of raw fatigue data and for environmental effects such
as the presence of a corrosive environment (e.g. salt water in winter). Also
required are robust and accurate algorithms for assessing the equivalent
fatigue life under complex dynamic loading conditions and multi-axial stress
states. Fortunately sophisticated fatigue analysis software to carry out this
complicated assessment is now available as part of established CAE systems or
as stand-alone packages.
Assuming that suitable fatigue data and analysis routines are available, the next
problem is to obtain appropriately detailed stress distributions for input to the
fatigue software. As discussed above, modern finite element methods are capable
of analysing stresses and strains in general 3D structures under almost any
complexity of loading conditions. The question is how complex does the analysis
have to be, as increasing complexity invariably means more expense and time
delay. Ideally one would like to make use of linear static stress analysis results but
there are limits to the applicability of such data, as discussed in the Case Study
below. It may also be possible to make use of a frequency domain approach to
account for the dynamic response of the structures considered. The alternative is to
carry out a full dynamic transient analysis for a typical time history of loading
input to the vehicle. By its very nature, this will be computationally intensive but
also by definition will pro- duce the most accurate results.
Whichever form of finite element stress analysis is undertaken, it is necessary
to have detailed knowledge of the input loads from the road surface or any other
source that give rise to the stresses in the individual components of the assembly
or structure. Such data can be generated from in-service or proving ground tests
but as discussed above, such tests are time-consuming and require a prototype vehicle
to be available. An alternative approach is to use multi-body dynamics (MBD)
simulation of the whole or part of the vehicle as it travels over typical road
surfaces in order to generate loading data at particularly critical locations.
Although MBD software is now readily available, many issues remain concerning
its accurate and effective use such as whether a quarter or full car model is required
and what type of tyre model to specify for each application.
Assuming that the above software elements are satisfactorily developed and
integrated so that accurate estimates of fatigue life distribution throughout the
structure/component can be made, there remains the possibility of using these life
distributions to automatically modify the designs leading to a more optimal
solution in terms of weight and/or efficient use of material. Although structural
optimisation routines do exist in some FEA packages, it is fair to say that they are
not yet routinely used in the automotive industry due to their limited applicability
to real structures subjected to dynamic loading conditions. This is particularly true
for the design of components and structures whose primary failure mode is likely
to be fatigue due to complex loading rather than from a single load application.
The extension of existing optimisation methods and, in particular, the powerful
 2 4 Vehicle Structures and

technique of Evolutionary Structural Optimisation (ESO) to this class of problem

is presented in the case study that follows.

4.5.3 Case Study: Durability Assessment and Optimisation

of Suspension Component

The component under consideration is the lower arm from the front suspension of
a multi-purpose vehicle as shown in Fig. 4.27. A more detailed view of the
existing design of arm is shown in Fig. 4.28. It can be seen that the arm has a
number of

1 2 3

4
y
z
5

x 6

1- torsion bar, 2-damper, 3- upper arm

4- bump stop, 5- lower arm, 6 tie rod, 7- hub

Fig. 4.27 Front suspension system of multi-purpose vehicle

Fig. 4.28 Lower suspension arm of multi-purpose vehicle

 4.5 Durability 2

(a) Quarter vehicle model showing simple spring representation of tyre in contact with road

(b) Full vehicle model

Fig. 4.29 Multi-body dynamics (MBD) models of multi-purpose vehicle

interfaces to other components of the suspension. In fact, a total of 19 external

load histories can be identified to act in the 3 co-ordinate directions at these
interfaces. In addition, there are 9 internal load histories due to the inertia of the
arm and its accelerations in 3D space to be considered. Thus the complete loading
on the arm is extremely complex and difficult to quantify.
In order to estimate these loads using a VPG approach, MBD models were
created and driven over a virtual version of a particular pavé durability track as
indicated in Fig. 4.29. Experimental results were also generated by driving an
actual vehicle with accelerometers attached at critical locations over the physical
pavé test track.
Although quarter vehicle models (QVM) were tried, Fig. 4.29a, it was found
necessary to allow for body roll/pitch by moving to a full vehicle model (FVM),
 2 4 Vehicle Structures and

Fig. 4.30 Power spectral density (PSD) plots of wheel hub accelerations

Fatigue life estimation

Quasi-static stress analysis Transient stress analysis Harmonic stress analysis

Fig. 4.31 Different fatigue life assessment strategies

Fig. 4.29b, in order to give accurate

Time domain predictions of measured
Time domain suspension
Frequency domain
life life life
component loading
assessment
histories. assessment assessment
Comparisons between the accelerations at the front wheel hub predicted by the full
vehicle model and the experimentally measured accelerations are shown in the Power
Spectral Density (PSD) plots of Fig. 4.30. It can be seen that both the amplitude
and frequency content of the experimental lateral and vertical accelerations of the hub
are well characterised by the FVM results which gave confidence that this model could
be used for the subsequent design optimisation studies.
Having established the external and internal inertial loading histories applied to
the component of interest (the lower suspension arm), the next stage is to carry out
a detailed 3D FE analysis of the arm to predict stresses for the subsequent fatigue
assessment. Here the user is faced with a choice of at least three different analysis
strategies as indicated in Fig. 4.31.
 4.5 Durability 2

Fig. 4.32 PSD of forcing function compared with arm natural frequency response

The quasi-static analysis strategy requires a single linear static analysis of the
arm for unit loading applied in each of the 19 external and 9 internal load
directions identified. Using the principle of superposition, the detailed stress
history distri- bution of the arm can then be obtained by factoring and summing
the influence coefficients generated by each analysis according to the actual
external load his- tories applied. Although this is the simplest and most
economical of the available methods, it was found to be accurate only if the
natural frequencies of the arm (which can be readily predicted from the FE model)
are well separated from the frequency content of the forcing function input, as
indicated in Fig. 4.32a.
If the frequencies are not separated by a factor of at least seven then the quasi-
static method was found to be highly erroneous because it does not account for the
interactions between the dynamic response of the component and the forcing
frequencies, potentially leading to resonance effects.
If the component and loading frequencies are separated by a factor of less than
seven as in Fig. 4.32b, then recourse must be made to the computationally
accurate but time consuming transient stress analysis method which requires full
integration and solution of the dynamic FE model at small time increments for the
complete loading time history considered. Such a computationally intensive
method is not suitable for the multiple simulations required for design
optimisation based on fatigue life even with the power of modern computer
systems.
Finally, if the forcing function and component natural frequencies overlap as in
Fig. 4.33c, it may be possible to use harmonic stress analysis techniques followed by
frequency domain life assessment to estimate the fatigue life distribution in the
component. Here, the arm transfer functions which give the stresses per unit load as a
function of frequency are firstly generated by applying either full or reduced
modal superposition techniques. The transfer functions are then multiplied by the
PSD’s of the applied loads to give the equivalent PSD’s of stress. Finally, the
fatigue life can be estimated in the frequency domain using the well-established
Dirlik approach.
When using the harmonic stress analysis method, there are a number of
parameters that have to be specified such as the number of modes to include in the
harmonic analysis, the buffer size used to perform the Fast Fourier Transform
 2 4 Vehicle Structures and

Evaluate load histories

Calculate stress histories

Estimate fatigue life for each element within the finite element model

Select and remove set of finite elements with high fatigue life

No
Objective achieved?

Yes
Stop

Fig. 4.33 Basic optimisation strategy based on fatigue life

(FFT) to generate the loading PSD’s and the frequency step size used to derive the
stress PSD’s. Although the values of these parameters have been shown to affect
the absolute fatigue lives predicted using this method, the distribution of fatigue
life
within the component is much less sensitive and is in good agreement with the
predictions of the most accurate transient stress analysis method. Since
optimisation of the design based on fatigue life depends on the distribution rather
than absolute values of life, the harmonic stress analysis method can therefore be
used for this purpose for cases where the forcing and natural frequencies
coincide as in Fig. 4.32c. This is far more computationally efficient than using
the full transient analysis method.
A basic optimisation strategy based on fatigue life is shown in Fig. 4.33. An
immediate problem arises when attempting to implement such a strategy using
standard S-N fatigue data in that a fatigue cut-off limit at around 10 8 cycles is
usually specified due to the lack of data for higher numbers of cycles. This means
that the majority of material in most automotive components will be predicted to
have infinite fatigue life. Since the optimisation algorithm seeks to remove
material with maximum life, a distribution which varies throughout the component
is really required so that the algorithm can start to remove material at the location
with the least damage. A practical solution to this problem is to artificially extend
the fatigue life cut-off point much further along the life axis, as indicated in Fig.
4.34. The fact that the fatigue curve beyond the standard cut-off limit may not be
completely accurate is irrelevant in terms of the optimisation strategy because
material with such very long predicted life will be rapidly removed from the
model.
To demonstrate the potential of this optimisation strategy to not only modify
existing designs but to generate radical new solutions, the maximum initial
domain of the lower suspension arm in question was established as a simple
rectangular
 4.5 Durability 2

Extended
cut-off limit

Fig. 4.34 S-N curve with artificially extended fatigue life cut-off limit

Fig. 4.35 Finite element model of initial domain for optimising lower arm

block, Fig. 4.35. Since it is necessary to have certain features (principally

mounting holes) at certain locations in the arm, these features were designated as
non-design domains which could not be altered by the optimisation routine.
Figure 4.36a shows that the initial fatigue assessment of the design domain
before modifying the fatigue cut-off limit predicts infinite life for most of the
material. In contrast, Fig. 4.36b shows the corresponding life contours after the
limit was raised to 1020 cycles.
Starting from the much more informative life distribution shown in Fig. 4.36b,
the optimisation routine was run with the objective of minimising the weight of
the arm subject to the constraint that at least 10,000 driving cycles over the
pavé
 2 4 Vehicle Structures and

(a) original fatigue cut-off limit

(b) extended fatigue cut-off limit

Fig. 4.36 Contour plots of fatigue life distribution in suspension arm

durability road should be possible without fatigue failure. Figure 4.37 shows the
history of material removal recorded during this optimisation process.
After about 80 loops of the algorithm shown in Fig. 4.33, the volume of the
arm was stabilised at less than 20% of its initial value with the resulting shape
and
 4.5 Durability 2

Fig. 4.37 History of material removal during optimisation of arm

Fig. 4.38 Contour plot of fatigue life distribution for optimised arm

fatigue distribution as shown in Fig. 4.38. The optimised arm has a fatigue life of
36,000 cycles of the pavé durability track compared to only 173 cycles for the
existing arm design shown in Fig. 4.27. Simultaneously the weight of the
optimised arm has been reduced by 0.5 kg compared with the original design.
Of course, the jagged surfaces created by the particular finite element mesh of
the optimised arm shown in Fig. 4.38 are neither desirable nor practical to
manufacture. Therefore the final stage of the process was to smooth these surfaces
to create the final optimised geometry shown in Fig. 4.39. Although somewhat more
complicated to manufacture, this optimised component uses significantly less
material than the original design shown in Fig. 4.28 and would have a fatigue life
even greater than
 2 4 Vehicle Structures and

Fig. 4.39 Final optimised design after surface smoothing

predicted for the “jagged” finite element model shown in Fig. 4.38 because of the
reduction of stress concentrations.
Although the optimisation process is unquestionably complex, the exercise
does indicate the potential of computerised tools and the Virtual Proving Ground
approach to not only optimise existing designs but also to generate novel solutions
such as shown in Fig. 4.38, which are unlikely to emerge from traditional design,
make and test practices. Such novel solutions also lend themselves to additive
manufacturing technologies which have fewer constraints on the part geometry
than conventional manufacturing processes.

4.6 Concluding Remarks

This chapter has covered some of the more important aspects concerning the
design and analysis of chassis structures. The consideration of alternative high
strength lightweight materials has been included because they have started to
revolutionise the ways that vehicle body structures are designed and fabricated.
Although advanced computer analysis methods have become the norm in the
industry for both static and dynamic assessment of structures (including their
durability and crash safety), it is important that the engineers understand the basic
principles and can, from time to time, apply conventional methods of analysis as a
reality check on computerised solutions.