AbstractThe objective of this paper is to study the analysis and

testing for determining the torsional stiffness of the student formulas

space frame. From past study, the space frame for Chulalongkorn
University Student Formula team used in 2011 TSAE Auto Challenge
Student Formula in Thailand was designed by considering required
mass and torsional stiffness based on the numerical method and
experimental method. The numerical result was compared with the
experimental results to verify the torsional stiffness of the space frame.
It can be seen from the large error of torsional stiffness of 2011 frame
that the experimental result can not verify by the numerical analysis
due to the different between the numerical model and experimental
setting. In this paper, the numerical analysis and experiment of the
same 2011 frame model is performed by improving the model setting.
The improvement of both numerical analysis and experiment are
discussed to confirm that the models from both methods are same.
After the frame was analyzed and tested, the results are compared to
verify the torsional stiffness of the frame. It can be concluded that the
improved analysis and experiments can used to verify the torsional
stiffness of the space frame.
KeywordsSpace Frame, Student Formula, Torsional Stiffness,
TSAE Auto Challenge
I. INTRODUCTION
SAE Auto Challenge Student Formula [1] is the student
formula competition of Thailand under formula SAE
international rules [2] organized by society of engineers
Thailand. In order to participate in this competition, the formula
car was formulated by considering many parts using the
engineering skills. The frame is one of the most important parts
that integrated the other parts together. The best frame is the
frame with less mass but still have high torsional rigidity. To
design frame, the analysis of torsional stiffness of the frame and
verify this results by using the torsional stiffness test became
more important [3][5].
There are many types of frame been used in Student Formula
Competition for example: Space Frame, Monocoque, and
Ladder Frame [6], [7]. A space frame, as shown in Fig. 1, is a
truss-like, lightweight rigid structure constructed from
interlocking struts in a geometric pattern. Space frames are a
series of tubes which are joined together to form a structure that
connects all of the necessary components together. Space frame
is simple and cheap when compared to the other twos. Therefore
it can be seen commonly in student level competition.
Tanawat Limwathanagura, Chartree Sithananun, and Teekayu
Limchamroon are with the Automotive Design and Manufacturing Engineering,
Chulalongkorn University, Phayathai, Pathumwan, Bangkok, 10330, Thailand.
Thanyarat Singhanart is with the Departments of Mechanical Engineering,
Chulalongkorn University, Phayathai, Pathumwan, Bangkok, 10330, Thailand
(e-mail: [email protected]).
Fig. 1 Space frame
In advance level of racing, Monocoque, as shown in Fig. 2, is
more popular. Monocoque is a construction technique that
supports structural load by using an objects exterior. It is
generally made as one piece. The use of composite materials in
monocoque skins allows strength, stiffness and flexibility to be
controlled in different directions. With the used of composite
material, Monocoque are significantly higher in cost but the
performance is worth investing. The last one isnt much seen, it
is the ladder frame. The ladder frame is a shorthand description
of a twin-rail chassis. It is typically made from round or
rectangular tubing similar to space frame. Ladder frame can use
straight or curved members, connected by two or more cross
members. Body mounts are usually integral outriggers from the
main rails, and suspension points can be well or poorly
integrated into the basic design.
Fig. 2 Monocoque
The space frame was chosen for 2011 Auto Challenge
Student Formula. From previous research [5], the torsional
stiffness of the 2011 frame was designed by considering
required mass and torsional stiffness with the usage of the
numerical and experimental methods. The numerical result of
Tanawat Limwathanagura, Chartree Sithananun, Teekayu Limchamroon, Thanyarat Singhanart
The Frame Analysis and Testing for Student Formula
T
World Academy of Science, Engineering and Technology 65 2012
1008
torsional stiffness was verified by using experiment. It can be
seen that the error between numerical result and experimental
result is large due to many mistakes when setting the
experiments. In this paper, the improvement of the numerical
model and experiment is the interested topic that has to be
carefully considered. Therefore, the objective of this paper is to
study the analysis and testing for determining the torsional
stiffness of the student formulas space frame.
II. THE ANALYSIS ANDEXPERIMENT OF 2011 FRAME [5]
The 2011 frame of Chulalongkorn University team for 2011
Auto Challenge Student Formula is shown in Fig. 3. The frame
was designed under SAE international rules with many concepts
such as
1) Lower the center of gravity by using the curve beam
2) Use the triangular mesh at the rear to improve the torsional
stiffness of the frame
3) Use removable member at the end to provide the space for
attaching the engine, etc.
Additionally, the frame was designed under the requirements
of mass less than 30kg and torsional stiffness of the whole car
more than 1200Nm/deg.
The Model for FEM analysis is shown in Fig. 4. The frame is
fixed at one end and the torque is applied at the other end. From
Fig. 4, the deflections at point 1, 2, 3, and 4, which located at the
rear part of the frame, are determined when fixed at the front.
The analysis was also performed when fixed at the rear and
torque is applied at the front to determine the deflections at
point 5, 6, 7, and 8, which located at the front part of the frame.
The torsional stiffness of the frame is from the lowest value
which occurred from point 7 when fixed at the rear. The
torsional stiffness from numerical analysis is found to be
1270Nm/deg.
Fig. 3 The 2011 frame for 2011-2012 Auto Challenge student formula
Fig. 4 Model for FEM analysis
The torsional stiffness test was performed as shown in Fig. 5
and 6. The steel rod is used to connect to the eight points at the
front which can be considered to be fixed end. At the other end,
the frame is connected to the steel beam that lied on the pivot
point. The load is applied at one end of the steel bar and
produced the torque to the frame as shown in Fig. 7. The torque
applied on the frame can be determined from
( )
1 2
T mgL F F w = = +
(1)
Where m is the end mass and L is the distance between end mass
and pivot point.
The deflections of point 1, 2, 3, and 4 as shown in Fig. 4 are
collected corresponding to the increasing load. The deflections
of point 5, 6, 7, and 8 can be determined with the rear part fixed
and torque applied to the front. The angle of rotation as shown
in Fig. 8 can be calculated from
1
tan
a b
L

=
(2)
The torsional stiffness from experiment can be determined
from
T =
. The experimental results show that the
torsional stiffness of the frame is 787Nm/deg. By comparing the
numerical and experimental results, the error is found to be
38.0% which is considered to be large.
The errors may come from many parts that will be discussed
in next section. After the source of error is identify, the
improvement of the experiments and numerical analysis can be
clearly obtained.
Fig. 5 Model for experiment setup
s:--. --a
c----:.-,
:- :-- a-
s:--. --a
-..-: ,-.-:
-.-+
--.-:s
/,,..-+
:-|a-
World Academy of Science, Engineering and Technology 65 2012
1009
(a)
(b)
Fig. 6 Experimental setup for torsional stiffness tests
Frame
Steel Beam
1
F
2
F
mg
L
2w
Fig. 7 Free body diagram of separated frame and steel beam
1
F
2
F
Frame

w
a
b

Fig. 8 The angle of twist
III. THE DISCUSSIONONTHE PAST RESEARCH
In the experiment, the connection points between the test rig
and the frame at the rear are joined by welding two steel
columns to the frame as shown in Fig. 5 or 6(a). It can ensure the
lock between the frame and the test rig however, caused by
welding these two members, the two part acting like one body.
As the result of this action, the point where load was distributed
into the frame has much error and fails to simulate the driven
condition. The process of holding the frame where the loads will
be applied should be improved to gain better result for the
torsional stiffness.
As can see in Fig. 5 and 6(a), the front part was fixed by
welding eight rods to frame. The torque applied will cause the
frame integrated with eight rods to rotate, then if the angle of
twist is calculated from the absolute deflections at the rear, this
value will represent the angle of twist of the whole system but
not only the frame. To ensure that only the frame rotates with
fixed support at the front, the improvement of the fixed points
has to be considered. It would be more accurate to ensure that
the fixed parts dont have any movement. The defection at the
fixed part must be measured and subtracted from the deflection
at the rotating end to get the displacement from frame rotation
only. Also, in measuring the deflection, suitable equipment
should be selected to measure distance a and b.
IV. THE IMPROVEDANALYSIS
In the finite element analysis the method for applying load is
changed as shown in Fig. 9. The force was distributed equally
into all four point of the frame on each side (point 1, 2, 3, 4)
when the rear was fixed at point 5, 6, 7, and 8. This model
verifies the condition to be the most related to the real
experiment.
Fig. 9 The improve analysis (The red points are the applied load points
while the purple are the construction points)

.z
.s

s:--. --a
-..-: ,-.-:
(/-,.- s:--.,
World Academy of Science, Engineering and Technology 65 2012
1010
V. THE IMPROVEDEXPERIMENT
For improving of the experiment first of all the way to
connect the frame and the test rig was changed. Instead of
welding the frame and the test rig together by a steel beam or
steel rods, new propose was introduced. In order to simulate the
closed situation in real driving, the force should distribute
through the frame passed by the wheel. Therefore not only the
frame, A-arms and wheel base were attached to the frame as
shown in Fig. 10. The new test rig is designed by connecting the
wheel base both at the front and at the rear to the test rig.
Fig. 10 The new test rig
This new test rig can be locked with wheel base by
combination of nuts and washers as shown in Fig. 11(a) and
11(b). The hold fitting object (Green part) minimize the
clearance between the hold and the wheel base. Due to the
inclination of the wheel base (Blue part) the different numbers
of washer (Red part) were added. Therefore the locker is
possible to position vertically with the test rig. After that
washers and nuts (Black part) are used to lock at the outer side.
This ensures the limitation of vertical movement and dont
allow the locker to slip out. One thing to consider is the number
of washer used. Washer usage is depending on how the wheel
base aligned with respect to the test rig. The front part of the test
rig can freely rotate by using the pin support as can see in Fig.
11(a).
(a) The locker
Wheel base
Washer
Hole Filling Object
Nut
Locker
(b) Side view of the locker
Fig. 11 Locker setting
For more accuracy of the experiment result the dial gauge and
vernier height gauge are used to measure deflection at the fixed
end and at the rotating end, respectively. The dial gauges with
0-25.0mm / 1.0 range and resolution 0.03mm/0.0012 are used
to measure the deflection at the fixed end. At the rotating end the
vernier height gauge will be installed to observe the deflection
at both sides. Actually dial gauges are used for checking
whether the fixed part has movement or not, if there is the
measurement gained, it will be subtracted from the rotating end
to ensure the exact displacement.
VI. RESULTS
The model when fixed at the front and the torque applied at
the rear was performed and the result from point 7 is used to
compare with past experiment as shown in Table I. The
improved experimental results for point 1, 2, 3, 4, 5, 6, 7, and 8
can be found to be 1371, 1104, 1251, 1060, 1143, 1483, 955,
and 1122, respectively. The results from the torsional stiffness
test and FEM analysis are presented in Table II. With these two
values of torsional stiffness the percentage of error can be
determined.
VII. DISCUSSION
From the previous experiment the error between the FEM
analysis and the experiment was 38.0% which is considered
being unacceptable and failing to verify the numerical result.
With the improvement presented in this paper, the result from
the improvement version has acceptable error between the FEM
analysis and the experiment. It shows a good way of improving
and proves that the improving of the analytical model and
experimental setting significantly affect the result. Therefore,
the numerical result is verified and can be used in design process.
However some change could be made to achieve more and make
both FEM analysis and exact test more precise.
FEM analysis can be used to predict the torsional stiffness in
order to design the new frame. The torsional stiffness of the
candidate frames based on the FEM analysis are compared and
selected for the best performance frame. However the exact
torsional stiffness test is required for the final frame design
anyway.
World Academy of Science, Engineering and Technology 65 2012
1011
VIII. CONCLUSIONS
The improved analysis and experiment are presented and
discussed in this paper. It can be seen that this improvement has
succeed to model the real situation of load applied to the frame
that can see from the small error between analytical and
experimental results. From now FEM analysis can be used as an
effective tool for designing a frame.
However lots of improvements are expected to be done in
order to enhance it to the next level. More designs are possible
to help improving this test rig to become more accurate and
flexible.
REFERENCES
[1] Auto Challenge Student Formula, Available at
www.tsae.or.th/autochallenge.html, Accessed on 20, Mar 2012
[2] Student Formula Rules and Regulations 2011, Available at
http://www.jsae.or.jp/formula/en/background.html, Accessed on 20, Mar
2012
[3] Abrams, R., Formula SAE Race Car Analysis: Simulation & Testing of
the Engine as a Structural Member, in FISITA 2008 World Automotive
Congress, Munich, Germany, 2008
[4] William B. Riley and Albert R. George, Design, Analysis and Testing of
a Formula SAE Car Chassis, in SAE International, Proceedings of the
2002 SAE Motorsports Engineering Conference and Exhibition,
Indianapolis, Indiana, 2002
[5] Sithananun C., Leelaphongprasut C., Baitiang C., Rungpipatphol N.,
Noomwongs N., Singhanart T., SAE Student Formula Space Frame
Design and Fabrication, in The second TSME International Conference
on Mechanical Engineering, Krabi, Thailand, 2011
[6] Edmund, F. Gaffney III and Anthony, R. Salinas, Introduction to
Formula SAE Suspension and Frame Design, SAE International, 1997
[7] Why Our Chassis Looks the Way It Does, Available at:
http://www.erareplicas.com/misc/stress/deslogic.html, Accessed on 20,
Mar 2012
TABLE I
COMPARISON BETWEEN IMPROVED AND PREVIOUS RESULTS

Torsional Stiffness [Nm/deg]
Comparison
FEM Experiment
Previous
Experiment
1270 780 38.0%
Improved
Experiment
1006 955 5.3%
TABLE II
IMPROVED TORSIONAL STIFFNESS FROM POINT 1-8
Point
Torsional Stiffness[Nm/deg]
Comparison
FEM Experiment
1 1371 1427 4.1%
2 1104 1157 4.8%
3 1251 1329 6.2%
4 1060 1129 6.5%
5 1143 1209 5.8%
6 1483 1547 4.3%
7 955 1006 5.3%
8 1122 1165 3.9%
World Academy of Science, Engineering and Technology 65 2012
1012