Application of FEA Techniques to a Hybrid Racing Car Chassis Design
Elliot Brinkworth, Daniel Jaggard, Martin Royds-Jones, Blake Siegler, David Barton, Andrew Deakin and Adam Heppell
The School of Mech. Eng., The University of Leeds, UK

Reprinted From: Proceedings of the 2000 SAE Motorsports Engineering Conference & Exposition (P-361)

Motorsports Engineering Conference & Exposition Dearborn, Michigan November 13-16, 2000
400 Commonwealth Drive, Warrendale, PA 15096-0001 U.S.A. Tel: (724) 776-4841 Fax: (724) 776-5760

The appearance of this ISSN code at the bottom of this page indicates SAE’s consent that copies of the paper may be made for personal or internal use of specific clients. This consent is given on the condition, however, that the copier pay a $7.00 per article copy fee through the Copyright Clearance Center, Inc. Operations Center, 222 Rosewood Drive, Danvers, MA 01923 for copying beyond that permitted by Sections 107 or 108 of the U.S. Copyright Law. This consent does not extend to other kinds of copying such as copying for general distribution, for advertising or promotional purposes, for creating new collective works, or for resale. SAE routinely stocks printed papers for a period of three years following date of publication. Direct your orders to SAE Customer Sales and Satisfaction Department. Quantity reprint rates can be obtained from the Customer Sales and Satisfaction Department. To request permission to reprint a technical paper or permission to use copyrighted SAE publications in other works, contact the SAE Publications Group.

All SAE papers, standards, and selected books are abstracted and indexed in the Global Mobility Database

No part of this publication may be reproduced in any form, in an electronic retrieval system or otherwise, without the prior written permission of the publisher. ISSN 0148-7191 Copyright 2000 Society of Automotive Engineers, Inc. Positions and opinions advanced in this paper are those of the author(s) and not necessarily those of SAE. The author is solely responsible for the content of the paper. A process is available by which discussions will be printed with the paper if it is published in SAE Transactions. For permission to publish this paper in full or in part, contact the SAE Publications Group. Persons wishing to submit papers to be considered for presentation or publication through SAE should send the manuscript or a 300 word abstract of a proposed manuscript to: Secretary, Engineering Meetings Board, SAE.

Printed in USA


Application of FEA Techniques to a Hybrid Racing Car Chassis Design
Elliot Brinkworth, Daniel Jaggard, Martin Royds-Jones, Blake Siegler, David Barton, Andrew Deakin and Adam Heppell
The School of Mech. Eng., The University of Leeds, UK

Copyright © 2000 Society of Automotive Engineers, Inc.

The Formula SAE and Formula Student competitions, held every year in the USA and UK, challenge teams of engineering students to design and build a small singleseater racing car. The University of Leeds has entered teams into these competitions for the past four years and has developed an award winning hybrid monocoque chassis design. The design enables a light, stiff and extremely safe chassis to be produced at a reasonable manufacturing cost. A chassis which is torsionally stiff enables a desirable roll moment distribution to be achieved for good handling balance. A chassis which can absorb high energy impacts whilst controlling the rate of deceleration will increase the likelihood of drivers surviving a crash without injury. This paper describes how Finite Element Analysis (FEA) techniques have been used to investigate both the torsional stiffness and crashworthiness of the chassis and how physical materials testing has been used to ensure the results are accurate.

The Leeds University entry for the 2000 (‘F4’ car, see Figure 1) Formula SAE and Formula Student competitions is a development of the 1999 car (‘F3’). Both cars use a hybrid monocoque design, which makes extensive use of composite sandwich panels. This design enables a light, stiff and safe chassis to be constructed without incurring the high tooling and labour costs associated with conventional composite designs.

Figure 1 - University of Leeds F4 (2000) car. While the team has only been involved in the competitions for a short time it has already gained several awards including the Formula SAE safety prize (both 1999 and 2000) and first and second place in the design event in 1999 and 2000 respectively. By applying finite element methods to the chassis design, the team is able to optimise the design of the car to produce a very light chassis while still maintaining the required levels of stiffness and crashworthiness.

The Formula SAE and Formula Student competitions are based around a design brief to produce a single seat ‘Formula’ style racing car suitable for the amateur racing driver. The car must be designed such that the unit cost is less than $9500 based on a production run of 1000 units per year. Cars compete in several dynamic events and are judged statically on a variety of criteria including design, cost and safety. The competition rules are kept as open as possible to encourage innovative design, but include a number of requirements to limit performance and ensure safety [1]

The hybrid monocoque chassis uses front and rear steel sub-frames connected by a composite structure. The sub-frames are used to distribute concentrated loads from the suspension and engine into the monocoque as well as providing the mandatory roll-over protection. Composite sidepods are bonded to the chassis to

provide side impact protection and space for items such as the radiator and fuel tank. A detachable fibreglass nosecone (changeable in under thirty seconds) containing a core of crushable foam satisfies frontal impact regulations. The nosecone and sidepods form sacrificial crush structures which prevent the main monocoque tub from being damaged and can easily be replaced. The monocoque tub, sidepods and associated bulkheads are constructed from sandwich panels using a ‘cut and fold’ technique. Several panel configurations are used in the construction, each with a different combination of core and facing materials. The welded steel sub-frames are bolted and bonded to the monocoque and act as jigs to hold the panel in place while the adhesive cures. The adhesive used is 3M’s DP490 2-part epoxy. Cuts made to the panel skins are reinforced using strips of carbon fibre prepreg bonded using Ciba Geigy’s Redux 322 2-part adhesive. The base of the sidepod forms a ground effect venturi that has been optimised using computational fluid dynamics (CFD) software to reduce lift at high speed. The front dampers and rockers are supported on a specially produced composite panel that has been laidup by the team. This uses a Nomex honeycomb core and 2-ply carbon skins. Concentrated loads are supported using Tufnol inserts, co-cured into the panel. The advantages of this method of construction are that the chassis is stiff, light and has excellent crashworthiness while the construction costs are favourable compared to other manufacture techniques. Spaceframes which are the norm at Formula SAE/Student require complex jigs and tooling to weld while the costs and labour associated with a prepreg composite lay-up are prohibitive. The chassis design of the 2000 car is an evolution of the 1999 design, the main differences being improvements in ergonomics to suit larger drivers, redesign of the rear sub-frame to shorten the wheelbase and redesign of the front damper mounting.

loading mechanism

Common Hypermesh model geometry, nodes, elements

material data, nosecone, masses (wheels, driver, engine)

Hypermesh Torsional Model

Hypermesh Crash Model

element types, material data, real constants, loading & constraints

Hypermesh ANSYS data interface

Hypermesh Dyna data interface

ANSYS Torsional Model

LS Dyna Crash Model

Linear Torsional Analysis

Crash Analysis

Figure 2 - Use of a common chassis model for both torsional and crash analysis. The common model uses shell elements to approximate the sandwich panels in preference to solid elements as they reduce the complexity of the model. The torsion model also uses 2D beam elements to represent the steel subframes for the same reason. The geometry of the engine and differential box have also been simplified. The engine is a solid block comprising first-order brick elements and the differential box is a hollow box made of first order shell elements. The HyperMesh model makes use of ‘collectors’ to group elements of similar materials or section properties, which allows the properties of each material to be easily altered. Previous work undertaken by Butler et al. [2] aimed to use a similar common model but there were complications with the HyperMesh – ANSYS link. To avoid these problems, a series of tests was undertaken to ensure that all of the required element types could be imported by ANSYS. As shown in Figure 2, the material and section properties for the torsional model are entered directly into ANSYS rather than being imported via the HyperMesh model (see Figure 3). This reduces the complexity of the data interface and allows the use of the multi-layer SHELL 91 element which is not directly supported by HyperMesh version 3.1.

Finite element analysis of the F4 chassis was undertaken as two separate studies, one to study the crashworthiness of the chassis and the other to optimise its torsional stiffness. Supporting this analysis was a programme of physical material tests to allow validation. Due to the differing nature of the analyses they were carried out using different solution processors, LSDYNA3D and ANSYS for crash and torsion analysis respectively. The aim from the outset was to develop a model of the car’s geometry that could be used by both packages. This was achieved using Altair’s HyperMesh pre-processor as shown in Figure 2. The use of a common model minimises the duplication of effort required to develop two separate models, as it can be customised to suit the two different loading conditions.

Figure 3 - HyperMesh finite element model of chassis.

Using Instrom tensile and compression test rigs, material tests were conducted on samples of the honeycomb sandwich panels, the two part polyurethane foam and the various fibre glass lay-ups. The following tests were conducted:

• • • •

Tensile tests on the honeycomb sandwich panels. Tensile tests on the fibre glass samples. Four point bend tests on the honeycomb sandwich panels. Compression test on the polyurethane foam. Figure 4 - Finite element model of a four-point bending test on a sandwich panel. Simple material models were also developed in HyperMesh to model the non-linear and failure properties of the panels. Parameters such as Young’s modulus, fracture stress and the maximum load allowable before bending occurs were defined for the sandwich panel. Again, the material tests detailed above were simulated in LS-DYNA3D using these models and compared to the results from the physical tests to confirm the validity of the material model. To model the crushable nosecone structure, the properties for both the fibreglass and the foam were derived from the material test data. The crush behaviour of the foam was modelled as a load curve of pressure versus volumetric strain. This was defined in HyperMesh, which allows the creation of load curves that can be applied to specific materials in order to simulate the actual behaviour of the material.

The data obtained from these experiments was used in the FE analyses conducted on the car as summarised in Table 1.
Material Test conducted Quantity of result Value Application to FEA analysis DYNA crashworthiness analysis ANSYS torsional simulation DYNA crashworthiness analysis DYNA crashworthine ss analysis

Honeycomb sandwich panel Honeycomb sandwich panel Honeycomb sandwich panel Polyurethane foam

Tensile test

Young’s modulus

E= 1.06 GPa

Four point bend test

Load curve


Four point bend test

Max load before failure Mass density and Bulk modulus

Pmax = 2.6 kN ρ= 50.7 3 kg/m and K= 0.63 GPa ---

Compression test

Polyurethane foam

Compression test

Load curve

DYNA crashworthiness analysis DYNA crashworthiness analysis

This aspect of the work employed linear finite element analysis to firstly determine the torsional stiffness of the F4 chassis and secondly to investigate possible design iterations to increase the stiffness of the chassis with little or no increase in mass. High torsional stiffness is a desirable attribute for a racing car design as it allows the handling of the car to be accurately controlled by varying the suspension parameters. A flexible chassis reduces the ability of the race engineer to translate a change in front to rear roll stiffness ratio into a change in front to rear load transfer distribution [3]. Thus the vehicle’s handling balance becomes more difficult to tune. Clearly any level of stiffness may be achieved when designing a chassis simply by adding more material to the more highly stressed areas of the structure. This is not good practice for a racing car, as extra weight will have a detrimental effect on its performance. For this reason, a better measure of a chassis’ torsional performance is the stiffness to weight ratio (specific stiffness), calculated by dividing the torsional stiffness by the structure mass.

Fibre glass samples

Tensile test

Young’s modulus and fracture stress

E= 7.29 GPa and σf = 231 MPa

Table 1 – Physical material test results and applications. Simple material models were developed in ANSYS to analyse the properties of the sandwich panels. These material models use multi-layer shell elements to allow the physical properties of each laminate to be specified individually rather than as the approximate properties of the corresponding monolithic material. The material tests detailed above were simulated in ANSYS using these models and compared to the results from the physical tests to confirm the validity of the material model, see Figure 4.

As described previously, SHELL 91 multi-layer elements were used to model the sandwich panel composites in the chassis. Using these material properties a baseline chassis model was analysed. The geometry of the model matches the ‘as built’ state of the F4 chassis. The chassis is loaded by applying equal and opposite forces, magnitude F, to the ends of a stiff rod, length d, that is rigidly attached to elements that occupy positions representative of the rocker mounts. The differential box has two nodes representing the rear rocker mounts constrained in all degrees of freedom except rotation about the global x axis, that is they are free to rotate about an axis parallel with the rear axle. Thompson et al [4] suggests that this arrangement is over-constrained and leads to artificially high stiffness predictions (approximately 9% too high). As the models are being assessed for comparative purposes only, this level of accuracy is acceptable. The stiffness is calculated by firstly recording the vertical displacements of the nodes at each end of the beam that is used to load the front cam mounts, UY1 and UY2. The angle of twist of the front end of the chassis is then calculated using the following formula:

The total combined front and rear roll stiffness is approximately 700 Nm/deg. The analysis conducted by Deakin [3] infers that, as the chassis stiffness is approximately equal to the total roll stiffness, 80% of the difference in roll stiffness exhibits itself as a difference in lateral load transfer between front and rear. Thus, compared to the F3 car (where the chassis was half as stiff as the roll stiffness) the suspension of the F4 car can be more readily tuned, with greater fine adjustment available. The baseline chassis finite element model was then modified to investigate the effect of a number of design iterations. The models for each iteration were analysed in the same manner and the percentage increases in stiffness and specific stiffness were compared to the baseline chassis. The results of this design study are shown in Figure 6.
50 40 30 % Change 20 10 0 -10 -20 -30
Seat bulkhead added Dash bulkhead added Sidepods Removed External dampers 2-ply carbon skins on sidepods Thicken honeycomb by 8mm Stiffening rib around cockpit 2-ply carbon outer skin

% Change in Stiffness % Change in Specific Stiffness

éUY + UY2 ù φ = arctan ê 1 ú d ë û


This is then used to calculate the torsional stiffness of the chassis.

Design Iteration

Kc =

F .d φ

Figure 6 – Results of different design iterations. (2) It can be seen from Figure 6 that the main opportunities for increasing the specific stiffness of the chassis are through increasing the thickness of the honeycomb cores and replacing the aluminium skins with carbon fibre. Further development in this area will yield a significantly stiffer yet lighter chassis for future cars.

The baseline chassis configuration shown in Figure 5 achieved a torsional stiffness of 880 Nm per degree from a structure weight of 19.69kg. This yields a stiffness to weight ratio of 44.6Nm/degree/kg. The overall stiffness is somewhat less than the 1000Nm/deg target identified but of the correct order of magnitude none the less.

This aspect of the work was concerned with the crashworthiness of the F4 chassis design. The main aim of the study was to analyse the behaviour of the chassis during two crash scenarios to ensure that the chassis structure would protect the driver during an impact. Racing cars are designed to be driven near the limit of adhesion at all times and are therefore prone to be involved in accidents, the most likely scenarios being a frontal or side impact. In order to validate the overall performance of the vehicle, both of these scenarios were simulated using the non-linear finite element analysis package, LS-DYNA3D.

Figure 5 - Vertical displacement plot of baseline chassis.

The competition rules [1] have a comprehensive section covering safety parameters, which specify the minimum requirements for a spaceframe chassis. Due to the composite construction of the Leeds chassis, energy absorbent structures are used to provide equivalent protection. These structures are the nosecone for frontal impacts and the sidepods for side impacts and are more than adequate when compared to the minimum requirements specified by the rules. The geometry of the chassis was defined in the common HyperMesh model described previously. However, as shown in Figure 2, the element attributes and boundary conditions still needed to be detailed in HyperMesh before the analysis could be run. It was found from previous work by Butler [2] that beam elements were inaccurate for modelling the rollover bars, so shell elements were applied. Rigid beam elements were used to constrain the junction points of the rollover bars as well as joining several other parts to the chassis. The HyperMesh package was also used to assign material properties and nodal masses to the model. The material modelling has been described previously. Masses can be assigned to individual nodes within HyperMesh in order to give the correct mass distribution. Table 2 lists how these masses were spread across the nodes representing each area of the car: COMPONENT Engine and ancillaries Differential box and driveline Wheel masses (x4) Chassis Driver TOTAL Table 2 - Distribution of the nodal masses. FRONTAL IMPACT - In frontal impacts, energy is absorbed in several ways. Firstly through the deformation of the nosecone and then through the deformation of the chassis itself. And finally through the movement of the driver and the stretching of the seatbelts. Due to time constraints and the high level of computing power required, it was impossible to run a simulation simultaneously involving both the chassis and nosecone as one integrated structure. The analyses undertaken simulated the behaviour of the nosecone and chassis when crashed separately. Also to simplify the analysis the driver was modelled as a collection of nodes rigidly attached to the chassis and positioned at the centre of gravity position assumed for the driver. MASS (kg) 85 40 20 each 40 75 320

The scenario simulated was of the impact of the car into a non-deformable object at approximately 13.4m/s (30 mph). An initial velocity of 13.4m/s was applied to all nodes within the model and a stationary rigid wall was created 5mm in front of the chassis. Previous work by O’Rourke [5] in conjunction with the Williams Formula 1 team and current FIA Formula 1 rules [6] quote suitable levels of deceleration that drivers can undergo without high levels of injury. The mean deceleration rate should be lower than 40g and the peak value should not exceed 60g for more than 3ms. The frontal impact simulation of the nosecone impacting the rigid wall, with the weight of the chassis behind it, enabled a force deflection curve to be generated. This showed how the nosecone limited the load transfer to the vehicle and the amount of energy (area under the graph) the nosecone absorbed. The frontal impact simulation of the chassis modelled the nosecone structure as four non-linear springs placed at the front of the chassis structure and directly contacting the wall. The performance of the nosecone was simulated using the force deflection curve from the nosecone impact analysis, which defined the behaviour of the non-linear springs. Figure 7 shows the average deceleration of all the nodes positioned at the driver’s centre of gravity. It can be seen that the average deceleration of the driver is 2 approximately 0.4 mm/(ms) (40g), with a peak for more 2 than 3ms of 0.75 mm/(ms) (75g). This is above the target set to avoid injury ([5] and [6]) and is due to two reasons. Firstly, that the restraint system and driver are not modelled. Secondly, using the F1 ([5] and [6]) regulations, the vehicle is simulated at a high impact speed. Due to the nature of the competition (low cornering speeds) and the track design (open car park with no rigid walls and track marked by cones), the speed of the vehicle would be much lower in an accident at this type of competition. As well, the formula SAE regulations [1] only require a relatively small crush zone compared with the F1 regulations ([5] and [6]). As a last point the chassis structure itself was not damaged showing that the nosecone acted as a sacrificial structure as planned.

Figure 7 – Average deceleration, of the driver nodes (in frontal direction), time history for frontal impact of chassis structure.

The LS-DYNA simulation has indicated that the majority of the impact energy will be dissipated in the nosecone before transmission to the chassis. The ability of the nosecone to absorb the impact energy was validated through a drop test on the F4 car’s spare nosecone. However, constraints on the drop test rig meant that an equivalent impact speed of only 8.9 m/s (20mph) had to be used. Figure 8 shows high speed camera frames from the drop test confirming that the deformation undergone by the nosecone is sufficient to absorb much of the impact energy.



Figure 9 – Picture of sidepod and rigid cylinder before the side impact simulation.



The torsional optimisation work to date has identified which design features present the most scope for improvement of both the torsional stiffness and specific stiffness of the chassis. Further work is required to identify the ideal configuration of the sandwich panels to be used in future chassis. Additionally, work is being undertaken to prove that economic construction of an all carbon skin chassis is viable.

Figure 8 – Nosecone drop test high speed camera pictures, in order of A to D. SIDE IMPACT – In a side impact collision the main consideration, with respect to driver injury, is cockpit intrusion. The main method of energy dissipation in this scenario is the deformation of the sidepods. The side impact simulation involved the vehicle impacting a non-deformable cylinder (simulating a pylon), again with a high impact speed of 13.4m/s (30 mph), see Figure 9. The simulation results showed that the cockpit suffers no intrusions due to the deformation of the sidepods which absorbed the impact energy and the stiffness of the cockpit sides. The average deceleration of the driver was found to be 30g and the peak value which lasted for more than 3ms was 57g. This is within the target set to avoid driver injury and again the driver and restraint system is not modelled. Also there was no damage to the monocoque itself, proving that the sidepod structure will act as a sacrificial structure as planned.

The authors acknowledge that the loading of the chassis to find its torsional stiffness is not a true representation of the torsional loading that the chassis experiences during use. To gain a better understanding of how the chassis stiffness effects the dynamic suspension performance, a modal analysis using the ANSYS models has been undertaken. This will be used in conjunction with an ADAMS multibody model of the car containing the suspension system, to study the effect the chassis stiffness has on the vehicle’s handling characteristics. In order to improve the work that has been carried out on the crashworthiness of the F4 car, several areas need to be looked at. Firstly, the amalgamation of the chassis with the nosecone model will help to improve the accuracy of the frontal impact simulation. Secondly all the crash analysis carried out to date has been concerned with collisions into rigid objects. In reality the car will impact with deformable objects [7], such as tyre barriers, cones or other cars. It would be more realistic to model these scenarios. Finally, the movement of the driver and stretch of the harness needs to be taken into account to fully determine if the driver would suffer an injury.

This paper has shown how finite element methods have been used by the Leeds University FSAE team to investigate the behaviour of a composite racing car chassis. The modelling tools and methods have been described including the development of a common model that can be used for both torsional and crash analyses. Material testing has been used to validate the material models giving confidence in the numerical results. The torsional analysis has yielded a stiffness figure of 880 Nm per degree which is close to the target. It has also identified means by which the torsional stiffness and specific stiffness could be further improved. The crash analysis has shown that the chassis can withstand high speed frontal and side impacts. Regrettably, simulation results show that high deceleration rates are imposed on the driver, which are close to threshold values and may cause injury. This is probably mainly due to the driver and restraint system model simplification. Areas for further work have been identified which should enable the team to develop even more competitive cars in the future.

1. SAE, Formula SAE Rules 2000, Published by SAE, SAE web-site. 2. Siegler B. P., Butler L., Deakin A. J., Barton D. C., The Application of Finite Element Analysis to Composite Racing Car Chassis Design, Sports Engineering (1999) 2 pp 245-252, September 1999. 3. Deakin A., Crolla D., Ramirez J. P., Hanley H., The Effects of Chassis Stiffness on Race Car Handling Balance, This Proceedings, 2000. 4. Thompson L. L., Raju S., Law E. H., Design of a Winston Cup Chassis for Torsional Stiffness, SAE Technical Paper Series 983053, 1998. 5. O’Rourke B., Chassis Design Workshop, Seminar material from Formula Student Workshop by Williams Grand Prix Engineering, held at University of Central England, November 1998. 6. F.I.A., F.I.A. web-site with Formula 1 technical regulations, F1teca.htm, August 2000. 7. Wright P.,Mellor A., Barrier Testing, SAE Technical Paper Series 983061, 1998.

Sign up to vote on this title
UsefulNot useful