STRUCTURAL ANALYSIS ON THE BLOODHOUND VEHICLE

Tim J Edwards, Gregor W Geddes

Atkins Aerospace

The Hub, 500 Aztec West

Bristol, United Kingdom, BS32 4RZ

ABSTRACT

Bloodhound is an extraordinary automobile. Designed for a top speed of 1050 mph, it is targeting an

attempt on the world land speed record in Northern Cape, South Africa in 2015 and 2016. Its design

speed makes it faster than a jet fighter at sea level and it is no surprise that the technologies

supporting its development are closer to aerospace than automotive.

The structure of the vehicle is a mix of metallic and carbon fiber composites and this paper examines

aspects of the vehicle structural integrity in the light of its mission profile, its configuration, its

structural arrangement, and the loads arising from its unique operational environment. The use of

composite stress methods that are closely aligned with those in the aerospace industry is highlighted,

the stress team having been led by engineers who have been involved in recent large aircraft

development programs. The paper concludes that the upcoming high-speed testing phase will provide

test data to support projected load and stress predictions.

1. INTRODUCTION

On 4th October, 1983, in the Black Rock Desert, Nevada, the British designed car Thrust 2, driven by

Richard Noble, recorded an average speed of 633.5 mph over two runs to set a new land speed

record. This vehicle was the first in a trio designed by a team led by Noble intended to break the

world land speed record. The latest incarnation, Bloodhound SSC, is designed to a mile-averaged

speed of 1000 mph, which, if achieved, would shatter the current record of 763.035 mph set up by its

forebear, Thrust SSC, in October 1997, and continue a 30-year succession of land speed records held

by the British team.

Figure 1: Thrust 2 and Thrust SSC

Figure 2: Bloodhound

Bloodhound’s structural environment is unique in terms of the loading cycle and the stresses that the

structure must withstand, and this report describes that loading cycle and the effects it has on two

structurally significant components:

1. The fuel tank enclosure and its attachments, which are highly loaded under fuel inertia;

2. The air brakes, which experience high aerodynamic drag loads when deployed at 800mph.

The design and analysis of these two items typifies the care and attention that has been applied to the

whole vehicle, in keeping with the aerospace technologies incorporated within it.

2. VEHICLE SPECIFICATION

Bloodhound’s mission is what sets it apart from other land vehicles: the car has no purpose other than

to go fast. However, there are constraints on its configuration dictated by design considerations such

as transportation to site, acceleration and braking performance and re-fuel time. Such constraints are

outside the boundaries of those that would affect typical vehicles. For example, in anticipation of the

trip to South Africa in 2014, a joint allows the 13.5m long bodyshell to be separated in two to allow

the removal of the Eurojet EJ200 jet engine, a pair of which also power the Mach 2.0 Eurofighter

Typhoon.

It is in regard to its extreme speed, however, that the greatest constraints on the car's configuration

are felt. To achieve a world speed record, the car must travel over a measured mile in opposite

directions, with the average speed over the two runs being recorded as the record speed. Further, the

two passes must be made within one hour of each other. As a consequence:

Track length is at a premium, as such speeds predicate a mile covered in around 3.6 seconds;

Re-fuel time is also a critical parameter in order to stay within the allotted time for the runs;

Both acceleration and deceleration are important, as time and distance cannot be wasted in

increasing and reducing speed.

Thus, powerful propulsion and braking are key elements of the performance of the Bloodhound

vehicle. Table 1 shows some performance parameters for Bloodhound compared with those for more

conventional vehicles. The very high power demand for Bloodhound is dictated by the need to travel

at speeds five times greater than a Ferrari 458, yet still to be accelerating at 0.5 g as it enters the

measured mile.

Table 1: Comparison of some Typical Parameters

Vehicle Weight (kg) Max Power

(hp)

Maximum

Speed (mph)

Average Fuel

Consumption

(mpg)

0-60mph (s)

Toyota Camry 1470 178 131 35.0 8.3

Ferrari 458 1400 605 205 20.0 2.9

Bloodhound ~7800 135,000 1050 0.04 ~10

2.1 General Structure of the Car

Bloodhound is of mixed material construction, comprising a carbon composite forward bodyshell

structure and a largely metallic aft bodyshell. With acceleration and deceleration phases of the

mission being so critical to success, light weight within the vehicle structure is vital, so the use of

composites in the forward structure, drawing on aerospace and motorsport technologies, is prevalent.

The aft section of the car requires metallic construction that can withstand the elevated temperatures

adjacent to the propulsion units. These two sections are joined at a mid-length joint, as shown in

Figure 3, which is used to allow the EJ200 engine to be removed.

Figure 3: Main Joint between Carbon Composite Forward Bodyshell and Metallic Aft Bodyshell

Construction techniques are strongly reminiscent of aerospace practice. The metallic aft

bodyshell is a monocoque fabrication utilising machined aluminium frames, or ribs, stiffening a

skin and longeron assembly fabricated in titanium, the parts being riveted together. The carbon

fiber reinforced plastic (CFRP) forward bodyshell utilises uni-directional (UD) and woven fabric

pre-preg material deployed in either monolithic laminates or as laminated face sheets on

aluminium honeycomb cores. The use of CFRP for the forward bodyshell allows the formation

of the complex curves necessary around the jet engine intake, whilst largely eliminating the

fasteners that would interrupt the smooth profile of the aerodynamic surfaces.

2.2 Mission Profile

Bloodhound’s mission lasts a little over two minutes, covering nearly twelve miles. In the

acceleration phase, which continues from the start until the peak speed is attained half-way through

the measured mile, the EJ-200 jet engine first propels the car to 350 mph at which speed a rocket

engine ignites, initially increasing acceleration to nearly 2 g, and continuing on up to the peak speed.

Throughout this acceleration phase the EJ200 continues to provide thrust.

Having achieved the peak speed and passed the measured mile, thrust from the jet and rocket engines

ceases and speed falls off rapidly as drag overcomes thrust. The car decelerates at 3 g and,

subsequently, at a decreasing rate, although progressive deployment of the air brakes maintains

deceleration at the maximum value possible until the car reaches a speed of 200 mph, at which point

wheel brakes can be applied.

Figure 4: Bloodhound Speed and Acceleration Mission Profile

These two peak acceleration and deceleration values determine longitudinal loads on the fuel which,

together with plenishment within the fuel tank and the vertical loads arising from unevenness in the

running surface, dictate the pressure loads generated within the fuel tanks. The structural analysis of

the tanks is discussed in Section 4. The air loads generated in decelerating the vehicle in the latter

half of its run are considered in the structural analysis of the air brakes, which are the subject of

Section 5.

Measured Mile

Vehicle Velocity

Figure 5: Position of Jet Fuel Storage and Airbrake Systems

4. ACCELERATION PHASE – JET FUEL SYSTEM

Bloodhound requires three different types of fuel when making a land speed record run. The hybrid

rocket engine uses a solid fuel within the rocket tube combined with a liquid oxidizer, the EJ200

engine runs off a jet fuel mixture and lastly the auxiliary power unit that drives the oxidiser pump

requires regular unleaded gasoline. The EJ200 is required to take Bloodhound from a standing start

to over 300 mph and then combine with the rocket engine to accelerate the car at a peak rate of 2 g to

over 1000 mph.

4.1 Jet Fuel System

The fuel system for the Rolls-Royce EJ200 jet engine comprises one main tank of 430 l capacity and

two auxiliary tanks of 80 l and 100 l capacity. Each tank features a Kevlar reinforced rubber bladder

enclosed within a sandwich panel constructed with CFRP face sheets and aluminium honeycomb

core. This is similar to the solution adopted by Formula 1 motorsport teams where the bladder is

encased within the composite monocoque structure of the car, immediately behind the driver.

Honeycomb sandwich construction is used also in fuselage fuel tanks in aerial tanker aircraft.

The shape of each enclosure, and indeed the need to have three separate tanks, is driven by the

available space inside the vehicle. An internal baffle was proposed for use within the main tank to

prevent fuel pooling at the front bulkhead during the deceleration phase and to reduce the movement

of the fuel throughout the run.

Jet Fuel Storage

Airbrake

4.2 Main Fuel Tank Analysis

Figure 6: Front & Rear Views of the Main Jet Fuel Tank

The fuel tank enclosure must be capable of withstanding the various combinations of acceleration

loading that Bloodhound may experience during a record attempt. A list of such combinations was

provided by the Bloodhound team for use in the analysis.

Table 2: Example Load Case Combinations

Case No. Forward

Acceleration (g)

Lateral

Acceleration (g)

Vertical

Acceleration (g)

1 - - 2.5

2 - 1.5 2.5

3 2.0 - 2.5

4 -3.0 - 2.5

5 2.0 1.5 2.5

6 -3.0 1.5 2.5

7 -9.0 - -

8 -9.0 1.5 -

As a final decision on the inclusion of the baffle had not been made in advance of the fuel tank

analysis, each loading combination was assessed twice; once with the baffle included and once

without it. Two primary load types were applied to the FEM for each load case; acceleration loading

on the enclosure itself (and any attached masses) and hydrostatic pressure of the fuel. The latter of

these was defined in the analysis using a number of equations to calculate the pressure at each

element of the enclosure surface by assuming the depth of the fluid was equal to the length of the

tank in the direction of loading. For load cases where the baffle was included it was assumed that the

fuel was fully partitioned by the baffle and therefore the pressure must be calculated separately for

the forward and rear sections of the tank. Note that for each load case the fuel tank was assumed to

be full in order to cover any eventuality where a run had to be aborted during its early stages.

Figure 7: Hydrostatic Pressure Distribution for a Deceleration Case (Baffle Included)

A mid-surface was extracted from a CAD model of the enclosure which was then idealised by

applying a mesh of 2D elements. A woven fabric CFRP was selected for the sandwich panels with

an isotropic layup and a cured ply thickness of 1.6mm per face sheet. 10 mm of aluminium

honeycomb was specified for the core of the panels.

Figure 8: Main Jet Fuel Tank FEM

The main fuel tank is attached at the front bulkhead to a metallic rail which runs from one side of the

vehicle chassis to the other. Additional hard points on the angled sides and at the rear of the tank

attach directly to the chassis. To minimise any load transfer from the chassis to the composite

enclosure, fittings with a degree of compliance were specified for these hard points.

Figure 9: Cross-Section of Compliant Fitting

These compliant mounts presented a significant challenge when modelling the joints in the FEM, as

no material had been specified for use in the joints and therefore the equivalent stiffness could not be

calculated. It was necessary also to consider that the design of the joint would mean that it would

have only a limited travel in any direction, after which the fastener would be pressing against the

hard point and the stiffness of the joint would significantly increase. To account for this the model

was run in four different configurations, using a different assumption for the stiffness of the

compliant joints each time.

1. Configuration 1: Assumed compliant material stiffness of 10 MPa

2. Configuration 2: Assumed compliant material stiffness of 100 MPa

3. Configuration 3: Assumed compliant material stiffness of 1000 MPa

4. Configuration 4: No compliance, joint stiffness calculated using material properties of

monolithic composite insert.

Five point masses, ranging from 5kg to 15kg, were attached to various points on the external surface

to simulate additional equipment being attached to the enclosure.

Figure 10 shows the distribution of stresses in the enclosure under 2.0g acceleration loading for

the four different model configurations discussed previously. Areas of elevated stress were

noted around each of the front rail attachments in configurations 1 (top left) and 2 (top right) but

were less significant in configurations 3 (bottom left) and 4 (bottom right). This pattern of

behaviour was expected as the decreased stiffness of the other attachment points in

configurations 1 and 2 results in the forward attachments reacting a greater proportion of the

applied loading.

4.3 Results

Figure 10: Von Mises Stress in the Enclosure for a 2.0 g Acceleration Case

A bolted joint analysis of each of the attachment points was carried out using techniques derived

from aerospace experience and focussing on the following failure modes:

Shear failure of the fastener

Tension failure of the fastener

Combined shear and tension failure of the fastener

Bearing failure of the composite face sheets

Shear failure of the threaded interface between the fastener and the threaded inserts

Shear failure at the interface between the monolithic insert and the honeycomb.

Fastener loads were taken directly from the spring elements representing them in the FEM. The

failure mode which gave the lowest margins of safety was the shear failure between the insert and the

honeycomb of the enclosure. This failure mechanism is driven by the through-thickness shear

strength of the honeycomb and the radius of the monolithic insert, as this dictates the area over which

the shear force can be distributed [2].

5. DECELERATION PHASE – AIR BRAKES

Upon exiting the measured mile, Bloodhound will immediately begin its deceleration phase. During

the record attempt it will be important that the vehicle can be stopped quickly, and in as short a

distance as possible, as it must be refuelled and ready to make the second run within 1 hour of the

first.

To achieve the required deceleration throughout the phase, three different braking systems will be

installed on the car.

1. Multi-plate Carbon Disc Brakes

2. Two Drag Parachutes

3. Two Airbrakes

As the throttles for the rocket and jet engines are closed off, drag forces acting on the car at 1000mph

cause an initial deceleration of 3.0 g (equivalent to 60-0 mph in under 1 second). As the car

approaches 800 mph the airbrakes are deployed gradually to maintain this deceleration. Once below

200 mph the wheel brakes are also used whilst the drag chutes are held in reserve.

5.1 Airbrake System

The airbrake system comprises two doors, each with a surface area of 0.6 m2. The doors are opened

using two actuators which are slaved together to prevent an actuator failure resulting in only one of

the doors opening.

Figure 11: Airbrake Mechanism

The perforations in the doors are designed to break up the airflow in such a way as to minimise

oscillatory loading on the rear suspension assembly, which is situated immediately behind the

airbrakes. Perforated airbrakes such as these are regularly seen in use on modern aircraft. The

positions of the actuator arm and the metallic hinges on each door are largely constrained by the

available space inside the vehicle, so the doors themselves can be no more than 50mm thick. To

avoid any possibility of the natural frequency of the doors coinciding with the frequency of vortex

shedding effects, a minimum first natural frequency of 45 Hz was specified. Any design must also

be capable of withstanding the aerodynamic loading when deployed whilst minimising the deflection

of the door, although no specific maximum value was specified. With these constraints in mind, a

design exercise was carried out to assess the most suitable materials to ensure the doors had the

required strength and stiffness characteristics.

5.2 Material Selection

Two top-level designs were considered throughout the study:

A door machined from a single piece of aluminium

A sandwich panel door, comprising aluminium honeycomb core and CFRP face sheets

The materials and properties tabulated in Table 3 were used throughout the analysis. For the CFRP

designs, both uni-directional and woven fabric materials were included. For all analyses the 0° fiber

direction is aligned perpendicular to the hinge line in the plane of the airbrake door. Nominal

stiffness values were assumed for the honeycomb core material.

Table 3: Airbrake Door Material Properties

M46J UD MTM49-3

(Uni Directional) [1] T700 2x2 Twill MTM49-3

(Woven Fabric) [1] Aluminium

E11 (MPa) 235000 57000 70000

E22 (MPa) 6870 54300 70000

G12 (MPa) 4090 3710 26000

S11T (MPa) 1709 842 469

S22T (MPa) 25.8 788 -

S12 (MPa) 69 73.3 283

5.3 Finite Element Analysis

A finite element model of the airbrake door was constructed by taking a mid-plane surface from the

airbrake door CAD model and creating a 2D element mesh on it. Mid-plane models were also

created in a similar way for each of the four hinge attachments to give a sufficiently accurate

representation of the stiffness of the assembly during modal analyses. The connections between the

hinges and the door were modeled using a combination of rigid body and spring element types. The

actuator connection was also modeled using a single rigid body element.

Figure 12: Airbrake FEM

Aerodynamic loading on the doors was taken from the results of CFD analysis of the rear section of

the vehicle for five different combinations of door position and velocity. To ensure a conservative

analysis, the airbrake is considered to start opening at 450 m/s (1000 mph) rather than 380 m/s (850

mph). The resulting loads were divided into three linear components, with Fx running from the hinge

line to the opposite edge of the door, Fy parallel to the hinge line, and Fz perpendicular to the plane of

the door. Under the assumption that any variation of air pressure across the door was negligible for

this study, each of the forces was then divided by the surface area of the door to give pressures, Nx,

Ny and Nz, which were applied in the analysis. These pressures are tabulated in Table 4.

Table 4: Aerodynamic Loading on Airbrake Door

v Fx Fy Fz Nx Ny Nz

mph kN kN kN kPa kPa kPa

Door Closed (0 °) 1050 0 0 11.0 0 0 19

Door Open (15 °) 1000 7.5 1.1 28.6 13 2 50

Door Open (30 °) 940 11.2 1.5 38.6 20 3 67

Door Open (45 °) 875 12.5 1.6 54.2 22 3 95

Door Open (60 °) 775 5.8 1.0 57.2 10 2 100

5.4 Results

A number of configurations were assessed for both the solid aluminium and sandwich panel designs.

Early analysis showed that in the sandwich panel designs the maximum deflection occurred at the

corner furthest from the actuator attachment. Since resistance to twist under static load was seen as a

measure of stiffness of a key deflection mode of the airbrake structure, efforts were made to reduce

this deflection and the layups under consideration were altered to include a greater number of ±45 °

plies.

Figure 13: Typical First Mode Displacement of a Sandwich Panel Design

Modal analyses were carried out with no aerodynamic loads applied using the Radioss solver. These

analyses showed that the first natural mode was torsional, with the trailing edge of the brake rotating

relative to the hinge line. Again, this mode shape encouraged the use of ±45 ° plies in CFRP designs

to increase the natural frequency in the most efficient manner. Eventually, four designs were settled

upon for comparison against the design criteria.

1. Single Piece Aluminium, 40 mm Thick

2. Sandwich Panel, 6 mm Uni-Directional Face Sheets with 38 mm Core

3. Sandwich Panel, 7 mm Woven Fabric Face Sheets with 36 mm Core

4. Sandwich Panel, 7 mm Uni-Directional & Woven Fabric Face Sheets with 38 mm Core

Figure 14: Von-Mises Strain for the Fully Deployed Case (Clockwise From Top Left: Designs 1 - 4)

Table 5: Summary of Airbrake Analysis Results

Description

1st Frequency

Mode of Door (Hz)

Maximum

Deflection (mm)

Maximum Von-

Mises Stain (µε)

Mass

(kg)

1 Aluminium 61 9.7 1080 70

2 Sandwich Panel,

UD Face Sheets 103 17.2 1550 18

3 Sandwich Panel,

Fabric Face Sheets 83 24.3 2440 19

4 Sandwich Panel,

Fabric and UD Face Sheets 96 17.8 1890 19

From the above results it was noted that, whilst the aluminium door had the lowest maximum

deflection, its first natural frequency of 61 Hz was significantly lower than the other designs. The

maximum von Mises stresses ranged from 86 – 105 MPa and Figure 14 shows that the load

distribution was similar in each design, with stress concentrations occurring at the edges of the holes

closest to the actuator arm. Using UD plies in the ±45° directions increased the first natural

frequency of the door as seen in designs 2 and 4. In the composite designs, Tsai-Hill [5] ply failure

analysis was utilised in the FEA solver to ensure that there were no predicted failures in the regions

of higher strain.

Based on these results the recommendation was made to use a composite sandwich panel design for

the airbrake doors, with the added incentive of these designs being estimated at less than 20 kg

versus 70 kg for the aluminium one. It was also concluded that optimisation of the ply layups and the

placement of the holes in the door to give more continuous fibres in the ±45 ° directions would offer

potential stiffness improvements. One way in which this could be achieved would be to use an

iterative finite element solver, which would be more efficient than the manual methods that had been

used hitherto.

Any detailed design would also have to be checked for global buckling and local failure modes such

as bearing / bypass failure around the hinge and actuator fasteners, face sheet wrinkling and local

compression failure of the honeycomb core; although preliminary calculations using methods drawn

from [3] indicate that the last two of these are unlikely to be significant.

6. CONCLUSIONS

The Bloodhound project is a further example of how composites are finding their way into many

areas of engineering. The strong association that the program has with aerospace and motorsport

technologies makes it an obvious application. The similarity of the Bloodhound composite

construction techniques to those used in aerospace, illustrated by the use of laminated pre-preg

raw materials and honeycomb sandwich panels, has allowed the use of analysis methods that can

be derived directly from aerospace experience. What the project does not do, of course, is

provide evidence of how composites can be applied in a large-scale, low-cost way into industries

that might benefit from the weight advantages that their use provides. This is the key challenge

for the future.

For Bloodhound itself, the paper describes just a fraction of the structural analysis that has been

performed to demonstrate the integrity of the vehicle for its mission. Added to this work is the

much greater engineering effort that has gone into the overall design of the car, reaching many

branches of discipline: aerodynamics, propulsion, structures and systems. The design analysis

has been supported by substantial testing of the propulsion system (especially the novel rocket

engine) and the aerodynamics (on an outdoor testing track). The project represents most aspects

of what would be seen in a major aerospace program. Nevertheless, Bloodhound’s mission is one

that moves further from what is known than ever before for ground speed record attempts, so the

risks are commensurately high.

Bloodhound has a robust heritage of record-breaking fore-runners, so we can be optimistic that it will

achieve its goal. Should Bloodhound be successful in its mission, it will represent the greatest

increase in the land speed record ever. Testing this year will increase confidence in the design and

the supporting analysis culminating in an assault on the record in 2015 and 2016.

7. REFERENCES AND ACKNOWLEDGEMENTS

1. ACG Motor Racing Database Mechanical Data Collation, Advanced Composites Group, 26th

Sep 2011

2. Failure Behaviour of Honeycomb Sandwich Corner Joints and Inserts, Sebastian Heimbs and

Marc Pein, 11th Dec 2008

3. HexWeb Honeycomb Sandwich Design Technology, Hexcel, Dec 2000

4. Bloodhound SSC Website, http://www.bloodhoundssc.com/

5. Jones R. M., Mechanics of Composite Materials, 2nd Edition, Taylor Francis, 1999, ISBN: 1-

56032-712-X