Beruflich Dokumente
Kultur Dokumente
II. OVERVIEW
The vehicle used for this study is been equipped with
MacPherson strut front suspension and twist beam rear
suspension. Various hardpoints have been modified in the
existing vehicle. The modified vehicle has been tested against
various test procedures and the test results have been
correlated with that of the mathematical model. For developing
the half car mathematical models ADAMS Car and for
developing full car model CarSim is been used.
MSC ADAMS[13] is a mechanical system dynamics
simulation tool widely used by suspension designers in
automotive industry. It is a virtual prototype software which
includes various interfaces for modelling, equation solving,
optimization, simulation and visualizing aids. It also enables
user to import rigid body models from different CAD
softwares and flexible bodies from packages like MSC Nastran
CarSim[14] developed by Mechanical Simulation Corp., is
used to simulate the vehicle dynamic behavior. It can
communicate very easily with softwares such as Matlab and
Simulink in order to run a co-simulation. Models typically run
more than ten times faster than real time on a 3 GHz PC,
enabling real-time (RT) testing with hardware in the loop
(HIL) systems.
The K&C characteristics of the optimized model from
ADAMS have been exported to CarSim where they are used to
build the full car model as it is easy to integrate vehicle Sfunction block with the controller model developed in
Simulink (As the vehicle would be equipped with autonomous
driving set-up which is not been discussed in this paper).
I. INTRODUCTION
http://dx.doi.org/10.15242/IIE.E0314507
66
2nd International Conference on Research in Science, Engineering and Technology (ICRSET2014), March 21-22, 2014 Dubai (UAE)
III. PROCEDURE
Firstly, for preparing half car models, suspension links are
modeled as rigid bodies, while, twist beam is modeled as a
flexible body imported from MSC Nastran. This is done in
order to ensure that the compliance characteristics of the
mathematical model closely resemble the actual vehicle.
Secondly, design guidelines and acceptance criteria of
vehicle suspension K&C characteristics are proposed. Thirdly,
key hardpoints which have great influence on K&C
characteristics are obtained by performing sensitivity analysis.
Hardpoint optimization study is been done on the selected set
of hardpoints to meet the desired objective function.
Finally, experimental correlation study is been performed
based on certain standard test procedures like ISO lane
change, Constant radius cornering, etc. to validate the
simulation results. The detailed procedure is shown below.
Fig. 2 Vehicle on jacks and half car front & rear Suspension models
67
2nd International Conference on Research in Science, Engineering and Technology (ICRSET2014), March 21-22, 2014 Dubai (UAE)
No Criteria Value
1 Camber 0to-0.5 deg
2 Caster 3 to 5 deg
0-0.3 deg
Toe-out
3 Toe
4 KPI
8 to 13 deg
5 Scrub -15to30 mm
6 CMA
20to30 mm
7 SAC
< 3 mm
No Criteria
Value
8 Ride Steer -6to-10deg/m
9 Ride camber -12to-18deg/m
10
11
12
13
14
Roll Centre
Anti-Dive
Anti-Lift
Susp Rate
Roll stiffness
110-150 mm
20 to 35 %
7 to 12 %
28-38 N/mm
1.2-2e3 Nm/deg
TABLE III
ACCEPTANCE CRITERIA: REAR SUSPENSION
No Criteria
1 Camber
2 Ride Steer
3 Ride camber
4 Roll centre
Value
No Criteria
0 to-0.5 deg
5 Anti-Dive
-6to-10 deg/m
6 Anti-Lift
-12to-18 d/m
7 Susp Rate
110-150 mm
8 Roll stiffness
Value
20 to 35 %
7 to 12 %
30-45 N/mm
.5-1.2e3Nm/d
Fig. 7 Rear Suspension with suggested modifications (encircled)
http://dx.doi.org/10.15242/IIE.E0314507
68
2nd International Conference on Research in Science, Engineering and Technology (ICRSET2014), March 21-22, 2014 Dubai (UAE)
TABLE V
FRONT SUSPENSION: OPTIMIZED DESIGN OBJECTIVES & HARDPOINTS
TABLE VI
REAR SUSPENSION: OPTIMIZED DESIGN OBJECTIVES & HARDPOINTS
The chart below shows the fit estimates for Toe angle as one
of the design variable for front suspension. Such fit estimates
were calculated for each of the design variable. It has been
verified for each of the design variable that it has a high F
value and low P value so that the model fitting is reliable. Note
that 128 runs were made, 1 run was for the first trial. The
model has 9 dof as we have 9 design variables, therefore, the
dof in which error can propagate is 118.
TABLE IV
MODEL FIT ESTIMATES FOR TOE ANGLE
VII. RESULTS
The set of optimized hardpoints obtained after Monte Carlo
optimization study are shown below.
http://dx.doi.org/10.15242/IIE.E0314507
69
2nd International Conference on Research in Science, Engineering and Technology (ICRSET2014), March 21-22, 2014 Dubai (UAE)
Acceptance Criteria:
1) Maximum Vertical Acceleration < 0.5 g
2) Max Pitch Rate < 15 deg/sec
3) Maximum pitch angle < 3.5 deg
4) Maximum H-point vertical change< 0 to 25mm
Boundary condition:
Speed Profile
http://dx.doi.org/10.15242/IIE.E0314507
70
2nd International Conference on Research in Science, Engineering and Technology (ICRSET2014), March 21-22, 2014 Dubai (UAE)
Acceptance Criteria:
1) Peak lateral acceleration < 0.6 g.
2) Susp roll gradient (Roll angle/ay)<4 deg/g
3) Body slip gradient (slip angle/ay)<5.5 deg/g
4) Pitch gradient (Pitch angle/ay)< 0.5 deg/g
5) Understeer gradient @0.5g< -30 deg/g
6) Understeer gradient @0.1g< -15 deg/g
Boundary conditions:
1) Initial velocity = 2m/sec.
2) Final velocity = Till vehicle reaches max lateral
acceleration,ay (constant ax of 0.1m/sec2)
3) Constant Radius tracks of 10,20,30 m respectively to find
understeer gradient variation with track radius
Results:
Results:
Results of all the above acceptance criteria are shown only for
20 m radius test. Moreover, variation of K is shown for tests
performed at all radius and at various levels of lateral
acceleration. The vehicle shows understeer characteristics as
indicated by the values of understeer gradient.
http://dx.doi.org/10.15242/IIE.E0314507
71
2nd International Conference on Research in Science, Engineering and Technology (ICRSET2014), March 21-22, 2014 Dubai (UAE)
ACKNOWLEDGMENT
The Authors would like to thank the Management of
Graduate School of IPS, Waseda University for providing the
financial support and allowing us to use the lab facilities of the
university. We would appreciate Onosokki Technical support
team who helped us a lot regarding the use of measurement
and acquisition system. We would also like to extend our
thanks to Toyota Body Company for making the vehicular
confidential data available to us specially tire related data
which are quite difficult to obtain and its determination
requires a lot of experimental effort.
REFERENCES
Results:
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
72
2nd International Conference on Research in Science, Engineering and Technology (ICRSET2014), March 21-22, 2014 Dubai (UAE)
[10] Habibi, H., K.H. Shirazi and M. Shishesaz, 2007. Roll steer
minimization of McPherson-strut suspension system using genetic
algorithm method. Mechanism & Machine Theory, 43(1): 57-67.
http://dx.doi.org/10.1016/j.mechmachtheory.2007.01.004
[11] Makita, M., 1999. An application of suspension kinematics for
intermediate level vehicle handling simulation. Society of Automobile
Eng. Japan, 20(4): 471-477
[12] Lee, S.H., U.K. Lee and C.S. Han, 2001. Enhancement of vehicle
handling characteristics by suspension kinematic control. Proceedings
of Institution of Mechanical Engineers, 215: 197-216.
http://dx.doi.org/10.1243/0954407011525584
[13] www.mscsoftware.com/product/adams
[14] www.carsim.com/products/carsim/
[15] www.onosokki.co.jp
[16] Thomas Gillespie, Fundamentals of Vehicle Dynamics
[17] William Milliken & Douglas Milliken, Race Car Vehicle Dynamics
[18] J C Dixon, Suspension Analysis and Computational Geometry
http://dx.doi.org/10.15242/IIE.E0314507
73