Optimal Design of Passenger Car Suspension For Ride and Road Holding
Optimal Design of Passenger Car Suspension For Ride and Road Holding
Optimal Design of Passenger Car Suspension For Ride and Road Holding
\
|
+
+ + + + +
+ + +
+ + + + + +
bF aF bF aF Yp Xp Z Zp XpCp
Yp Xp Z Zp XpKp Z W b Z bC Z W b Z bK
Z W a Z aC Z W a Z aK Z W b Z bC
Z W b Z bK Z W a Z aC Z W a Z aK Iy
(4)
( ) 0 1 1 1 ) 1 ( 1 ) 1 ( 1 1 1 = + + + +
F Q Z Kt Z W a Z C Z W a Z K Z M (5)
( ) 0 2 2 2 ) 2 ( 2 ) 2 ( 2 2 2 = + + + + + +
F Q Z Kt Z W b Z C Z W b Z K Z M (6)
( ) 0 3 3 3 ) 3 ( 3 ) 3 ( 3 3 3 = + +
F Q Z Kt Z W a Z C Z W a Z K Z M (7)
( ) 0 4 4 4 ) 4 ( 4 ) 4 ( 4 4 4 = + + + +
F Q Z Kt Z W b Z C Z W b Z K Z M (8)
Using following state space variables,
1 X Zp = 2 X Zp =
3 X Z = 4 X Z =
5 X = 6 X =
7 X = 8 X =
9 1 X Z = 10 1 X Z =
11 2 X Z = 12 2 X Z =
13 3 X Z = 14 3 X Z =
15 4 X Z = 16 4 X Z =
Substituting above variables in Eq.(1-8) and writing the
equations in state space representation form,
GF BQ AX X + + =
(9)
Where,
A16] A15 A14 A13 A12 A11 A10 A9 A8 A7 A6 A5 A4 A3 A2 1 [
T
A A=
16] 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 [
T
X X X X X X X X X X X X X X X X X =
T
B
B
B
B
B
(
(
(
(
(
=
4
3
2
1
(
(
(
(
(
=
4
3
2
1
Q
Q
Q
Q
Q
T
G
G
G
G
G
(
(
(
(
(
=
4
3
2
1
(
(
(
(
(
=
4
3
2
1
F
F
F
F
F
[ ]
T
B 0 0 0 0 0 0 (Kt/M1) 0 0 0 0 0 0 0 0 0 1 =
[ ]
T
B 0 0 0 0 (Kt/M2) 0 0 0 0 0 0 0 0 0 0 0 2 =
[ ]
T
B 0 0 (Kt/M3) 0 0 0 0 0 0 0 0 0 0 0 0 0 3 =
[ ]
T
B (Kt/M4) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 4 =
T
0] 0 0 0 0 0 1/M1 - 0 a/Iy - 0 W/Ix 0 1/M 0 0 [0 = G1
T
0] 0 0 0 1/M2 - 0 0 0 b/Iy 0 W/Ix 0 1/M 0 0 [0 = G2
T
0] 0 1/M3 - 0 0 0 0 0 a/Iy - 0 W/Ix - 0 1/M 0 0 [0 = G3
T
1/M4] - 0 0 0 0 0 0 0 b/Iy 0 W/Ix - 0 1/M 0 0 [0 = G4
0] 0 0 0 0 0 0 0 0 0 0 0 0 0 1 [0 = A1
0] 0 0 0 0 0 0 0 0 0 0 0 1 0 0 [0 = A3
0] 0 0 0 0 0 0 0 0 0 1 0 0 0 0 [0 = A5
0] 0 0 0 0 0 0 0 1 0 0 0 0 0 0 [0 = A7
0] 0 0 0 0 0 1 0 0 0 0 0 0 0 0 [0 = A9
0] 0 0 0 1 0 0 0 0 0 0 0 0 0 0 [0 = A11
0] 0 1 0 0 0 0 0 0 0 0 0 0 0 0 [0 = A13
1] 0 0 0 0 0 0 0 0 0 0 0 0 0 0 [0 = A15
T
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
|
|
\
|
0
0
0
0
0
0
0
0
XpCp -
XpKp -
YpCp -
YpKp -
Cp -
Kp -
Cp
Kp
Mp
1 -
= A2
T
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
\
|
C4 -
K4 -
C3 -
K3 -
C2 -
K2 -
C1 -
K1 -
XpCp) + bC4 + aC3 - bC2 + (-aC1
XpKp) + bK4 + aK3 - bK2 + (-aK1
YpCp) + C4) - C3 - C2 + (W(C1
YpKp) + K4) - K3 - K2 + (W(K1
Cp) + C4 + C3 + C2 + (C1
Kp) + K4 + K3 + K2 + (K1
Cp -
Kp -
M
1 -
= A4
Optimal Design of Passenger Car Suspension for ...
J. of the Braz. Soc. of Mech. Sci. & Eng. Copyright 2008 by ABCM January-March 2008, Vol. XXX, No. 1 / 69
T
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
\
|
C4W
K4W
C3W
K3W
C2W -
K2W -
C1W -
K1W -
XpYpCp) - bWC4 - aC3W + bC2W + (-aC1W
XpYpKp) - bWK4 - aK3W + bK2W + (-aK1W
Cp) Yp - C4) + C3 + C2 + (C1 (W
Kp) Yp - K4) + K3 + K2 + (K1 (W
YpCp) + C4) - C3 - C2 + (W(C1
YpKp + K4) - K3 - K2 + (W(K1
CpYp
KpY
Ix
1 -
= A6
2 2
2 2
T
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
|
|
\
|
bC4 -
bK4 -
aC3
aK3
bC2 -
bK2 -
aC1
aK1
Cp) Xp - C4) + C3 + C2 + (C1 (a
Kp) Xp - K4) + K3 + K2 + (K1 (a
XpYpCp) - bWC4 - aC3W + bC2W + (-aC1W
XpYpKp) - bWK4 - aK3W + bK2W + (-aK1W
XpCp) - bC4 + aC3 - bC2 + (-aC1
XpKp) - bK4 + aK3 - bK2 + (-aK1
CpXp
KpXp
Iy
1 -
= A8
2 2
2 2
T
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
|
|
\
|
0
0
0
0
0
0
C1
K1) + (Kt
aC1
aK1
C1W
K1W -
C1 -
K1 -
0
0
M1
1 -
= A10
T
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
\
|
0
0
0
0
C2
Kt) + (K2
0
0
bC2 -
bK2 -
C1W -
K2W -
C2 -
K2 -
0
0
M2
1 -
= A12
T
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
\
|
0
0
C3
Kt) + (K3
0
0
0
0
aC3
aK3
C3W
K3W
C3 -
K3 -
0
0
M3
1 -
= A14
T
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
\
|
C4
Kt) + (K4
0
0
0
0
0
0
bC4 -
bK4 -
C4W
K4W
C4 -
K4 -
0
0
M4
1 -
= A16
Active Suspension System
The linear time invariant system (LTI) is described by Eq.(9).
For controller design it is assumed that all the states are available
and also could be measured exactly. First of all let us consider a
state variable feedback regulator (Ogata, 1996);
KX F = (10)
Where K is the state feedback gain matrix.
The optimization procedure consists of determining the control
input F which minimizes the performance index. The performance
index J represents the performance characteristic requirement as
well as the controller input limitations. In this work LQR control
scheme is used to find the control force required, for which one has
evaluate the performance index J and hence design the optimal LQR
controller. The optimization procedure consists of determining the
control input F, which minimizes J, the performance characteristic
requirement as well as the controller input limitations.
+ =
0
dt ) F R F X P X ( J
T T
(11)
Where
16] 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1
T
X X X X X X X X X X X X X X X X [ X =
also
(
(
(
(
(
=
4
3
2
1
F
F
F
F
F and P and R are positive and are called weighting
matrices.
The function inside the integral in Eq.(11) is a quadratic form
and the matrices P and R are usually symmetric. It is assumed that R
is positive definite and P is positive semi definite. If R is very large
relative to P, which implies that the control energy is penalized
heavily, the control effort will diminish at the expense of larger
values for the state. When P is very large relative to R, which
implies that the state is penalized heavily, the control effort rises to
reduce the state, resulting in a damped system. P and R represent
respective weights on different states and control channels
respectively and are assumed accordingly.
Several procedures are available to solve the LQR problem. One
approach to find a controller that minimizes the LQR cost function
is based on finding the positive-definite solution of the following
Anil Shirahatt et al
70 / Vol. XXX, No. 1, January-March 2008 ABCM
Algebraic Riccati Equation. Linear optimal control theory provides
the solution of Eq.(11) in terms of Eq.(10).
The gain matrix K is computed from;
E G R K
T 1
= (12)
Where the matrix E is evaluated being the solution of the
Algebraic Riccati Equation ;
0
1
= + +
P E G EGR E A AE
T T
(13)
And substituting gain matrix K in eqn. 9 we get
BQ X GK A X + =
) ( (14)
While designing the LQR controller more weightage is given to
ride comfort and an upper limit of 25N is kept to the controlling
force depending on the design constraints to reduce cost function.
Due to their effectiveness in searching optimal design parameters
and obtaining globally optimal solution, the Genetic Algorithms are
applied to find the optimal actuator configuration.
Passive Suspension System
For passive suspension system as there is no actuator force i.e.
[F] =0 and Eq.(9) becomes
BQ AX X + =
(15)
The Eq.(14) and Eq.(15) can be solved for frequency domain or
time domain using Matlab [Ogata, 1996].
Optimization and Analysis
Analysis of the suspension system generally implies solving
Eq.(1-8) for the time response of the system. The following
optimization methods and procedure is adopted for analysis.
Optimization Problem Formualtion
The performance characteristics which are of most interest when
designing the vehicle suspension are passenger ride comfort, road
holding and suspension travel. The passenger ride comfort is related
to passenger acceleration, suspension travel is related to relative
distance between the unsprung mass and sprung mass and road
handling is related to the tyre displacement.
Among the above three characteristics ride comfort is chosen to
be the most important characteristic and is expressed in an objective
function as
) t ( Zp RMS ) Z ( f min
=
As per ISO2631 standards the passenger feels highly
comfortable if the weighted RMS acceleration is below 0.315 m/s
2
(Wong, 1998, Griffin, 2003 and ISO: 2631-1-1997). So, it is
considered as constraint.
0 315 0
2
1
= s / m . f g
At least 5 inches of suspension travel must be available in order
to absorb a bump acceleration of one-half g without hitting the
suspension stops and also an upper bound to maximum acceleration
should be kept so that at any time suspension will not hit suspension
stops (Baumal et al., 1998 and Gillespie, 2003). Both these are taken
as constraints
0 127 0 1
2
= m . Z Z g , 0 127 0 2
3
= m . Z Z g
0 127 0 3
4
= m . Z Z g ,
0 127 0 4
5
= m . Z Z g , 0 5 4
2
6
=
s / m . ) t ( Zp max g
Dynamic tyre force will increases with increase in tyre
deflection so an upper bound to maximum tyre deflection is placed
and it is considered as one more constraint (Baumal et al., 1998 and
Gillespie, 2003).
0 0508 0 1 1
7
= m . Q Z g , 0 0508 0 2 2
8
= m . Q Z g
0 0508 0 3 3
9
= m . Q Z g , 0 0508 0 4 4
10
= m . Q Z g
The other performance characteristic viz. road holding is
included as constraints and is restricted by (Baumal et al., 1998).
0 07 0 1
11
= m . Z g , 0 07 0 2
12
= m . Z g
0 07 0 3
13
= m . Z g , 0 07 0 4
14
= m . Z g
Human being feel comfortable within a frequency zone of 0.8
Hz and 1.5 Hz and also another criterion for good suspension
system often considered is the maximum allowable jerk experienced
by the passengers. Both these are added as two more constraints
(Griffin, 2003 and Gillespie, 2003).
Hz . Wn . g 5 1 8 0
15
= , 0 18
3
16
=
s / m ) t ( Zp Max g
In order to make pitch motion die faster natural frequency of
front suspension should be greater than the rear suspension and it is
considered as constraint (Gillespie, 2003).
Wr Wf g > =
17
Table 1 (Panzade, 2005) gives the details of fixed parameters
used in the analysis and the design variables are also restricted to
ranges defined by the bounds as shown in table 2 (Panzade, 2005).
Table 1. Fixed parameters
Parameters Values Parameters Values
Kt 200000 N/m Iy 4140 kg-m
2
Mp 100 kg 2W 1.450 m
M 2160 kg a 1.524 m
M1, M3 85 kg b 1.156 m
M2, M4 60 kg Xp 0.234 m
Ix 946 kg-m
2
Yp 0.375 m
Table 2. Variable design parameter ranges,
Design Parameters Lower bound Upper bound
Kp (N/m) 90000 N/m 120000 N/m
Cp (Ns/m) 400 Ns/m 900 Ns/m
K1, K3 (N/m) 75000 N/m 100000 N/m
C1, C3 (Ns/m) 875 Ns/m 3000 Ns/m
K2, K4 (N/m) 32000 N/m 70000 N/m
C2, C4 (Ns/m) 875 Ns/m 3000 Ns/m
Optimal Design of Passenger Car Suspension for ...
J. of the Braz. Soc. of Mech. Sci. & Eng. Copyright 2008 by ABCM January-March 2008, Vol. XXX, No. 1 / 71
Road Profile
A sinusoidal shape of the road profile as shown in Fig. 2
consisting of two successive depressions of depth h = 0.05 m, length
= 20 m and vehicle velocity V = 20 m/s is used for analysis
(Baumal et al., 1998).
As a function of time, the road conditions are given by
=
0
2
0 )), cos( 1 (
2 ) (
3 , 1
Otherwise
V
t if wt
h
t Q
and
\
|
+
=
0
2
)), ( cos( 1 (
2 ) (
4 , 2
Otherwise
V
tau t tau if tau t w
h
t Q
Where tau and w are the time lag between front and rear wheels
and the forcing frequency respectively and are given by
|
\
| +
=
V
b a
tau and
V
w
2
=
In this study, the right and left sides have same amplitude road
profile but there is a time delay of 0.2 sec and also the rear wheel
will follows the same trajectory as the front wheels with a time
delay of tau as shown in Fig.2. This road input will help to introduce
bounce, pitch and roll motion simultaneously.
0 0.5 1 1.5 2 2.5 3
0
0.01
0.02
0.03
0.04
0.05
0.06
Time (sec)
A
m
p
l
i
t
u
d
e
( m
)
Front Left
Rear Left
Front Right
Rear Right
Figure 2. Road profile.
Modified Objective Function
The constrained optimization problem is converted into
unconstrained one using penalty approach. The modified objective
function is stated as
c
G f Y + = (16)
Where f is the initial objective function and G
c
is a penalty when
constraints are violated and is given as
=
=
17
1 i
) , 0 max(
i c
g G (17)
In Eq.(17), is a penalty value which will vary between 8000
and 10000. A GA program is written in MatLAB, which initialize
suspension design variables. Then these values are passed into the
8DOF full car model to solve for the dynamic response of the
system. These values are then substituted back into the GA process
to calculate the fitness of the suspension design. This procedure is
repeated until the stopping criterion is met.
Results and Discussions
This section is divided into two parts. The first gives the best
parameters for the present models and comparison the results with
simulated annealing method while the second part deals with
simulation of present optimally designed suspensions.
The design results from the GA program for passive and active
suspension are tabulated in table 3 In order to verify the validity of
the results; the GA results were compared to those obtained by
simulated annealing technique.
Simulation is performed using vehicle data illustrated in table 1,
for road input defined in Fig. 2 and the optimal suspension
parameters defined in table 3 using genetic algorithm. In table 3 it
can seen the natural frequency of seat for both suspension systems
are within the comfortable zone of 0.8-1.5 Hz, while passive
suspension system the natural frequency is more compared to active
since it use more stiffer suspension at front and rear.
In Fig. 3 it can be observed that the reduction of the drivers
vertical displacement peak is approximately 74.2% in case of active
suspension as compared with passive suspension and also settling
time is reduced from 6 sec to 3.5 sec. Also it is observed that sprung
mass vertical displacement is less in case of active suspension
compare to passive suspension while pitch and roll displacement is
amplified in active suspension and, consequently return to zero is
also fast (Fig.4-6). This will occurs since during LQR controller
design for active suspension more weightage is given to vertical
displacement for comfortable ride.
From Fig. 7-8 it can be observed that seat acceleration and
sprung mass vertical acceleration is reduced by 88.72% and 88.17%
respectively in case of active suspension as compared with passive
suspension and also settling time is reduced from 6.5 sec to 3 sec.
Also the vertical weighted RMS acceleration of seat and sprung
mass is reduced from 0.3032 m/s
2
to 0.0534 m/s
2
and 0.2834 m/s
2
to
0.0492 m/s
2
since in case of active LQR controller design more
weightage is given ride comfort. It can also be observed sprung
mass weighted RMS acceleration is less than seat since seat is
located near front right side of tyre while for sprung mass weighted
RMS acceleration is calculated at center of gravity of sprung mass.
From Fig. 9 it can be observed that range of the roll acceleration
is 65% lower with active suspension than passive suspension. One
should be remind here that this rolling motion is excited by time
delay between the left and right side bump. Hence, the active
suspension has proved to be definitely superior to the passive case.
With regards to pitch acceleration illustrated in Fig. 10, for
active suspension the acceleration amplitude range is lower and
consequently, returns to zero is very fast. In addition, disturbances
of higher amplitude were recorded at about 0.6 sec and 1.6 sec. If
we analyze the excitation in Fig. 2 one can observe that these
disturbances are likely due to the phase angle of wheel motion
slightly ahead of disturbance.
From Fig. 11-14 it can be observed that in case of active
suspension system suspension travel increases by 56-60% than
passive suspension to provide more ride comfort i.e. less
displacement of sprung mass. Also tyre displacement is
approximately 28.5% less in case of active suspension than passive
suspension system, yielding better road holding (Fig. 15-18). Also it
can be concluded since suspension travel and road holding are
mutually contradicting parameters, there is increase in suspension
travel in case of active suspension than passive suspension.
Anil Shirahatt et al
72 / Vol. XXX, No. 1, January-March 2008 ABCM
Also from Fig. 19 and 20 it can be observed that rate of change
of acceleration is less in case of active suspension. Hence, jerk
experienced by driver seat and sprung mass, for active suspension
jerk is very less compare to passive suspension. Figure 21-24 gives
the actuator forces required for active suspension and all are well
below the applied limits and practically implementable.
Table 3. Design results of genetic algorithm and simulated annealing method.
Genetic
Algorithm
Simulated Annealing
Parameters
Passive Active Passive Active
Kp- Seat (N/m) 98935 95161 98946 95168
Cp Seat (N-s/m) 615 415 596 415
K1- Front Left (N/m) 96861 78158 96830 78098
C1 Front Left (N-s/m) 2460 2012 2458 2021
K2 Rear Left (N/m) 52310 41731 52331 41698
C2 - Rear Left (N-s/m) 2281 1848 2281 1863
K3 - Front Right (N/m) 96861 78158 96830 78098
C3 - Front Right (N-s/m) 2460 2012 2458 2021
K4 - Rear Right (N/m) 52310 41731 52331 41698
C4 - Rear Right (N-s/m) 2281 1848 2281 1863
RMS vertical acceleration of seat (m/s
2
)
0.3032 0.0534 0.3032 0.0537
RMS vertical acceleration of sprung mass (m/s
2
)
0.2834 0.0492 0.2833 0.0495
Max seat acceleration (m/s
2
)
2.0849 0.2350 2.0852 0.2357
Max. sprung mass acceleration (m/s
2
)
1.9172 0.2268 1.9175 0.2274
Max. seat displacement (m) 0.0725 0.0187 0.0725 0.0187
Max. sprung mass displacement (m) 0.0690 0.0181 0.0690 0.0182
Max. pitch displacement (degrees) 0.0222 0.0025 0.0222 0.0025
Max pitch acceleration (rad/s
2
) 1.1700 0.0582 1.1702 0.0583
Max. roll displacement (degrees) 0.0096 0.0029 0.0096 0.0029
Max roll acceleration (rad/s
2
) 0.5041 0.0888 0.5041 0.0890
Max. Suspension travel (m) Front Left side 0.0383 0.0320 0.0383 0.0320
Max. Suspension travel (m) Front Right side 0.0122 0.0305 0.0122 0.0305
Max. Suspension travel (m) Rear Left side 0.0290 0.0293 0.0290 0.0292
Max. Suspension travel (m) Rear Right side 0.0125 0.0288 0.0125 0.0287
Max. Road holding (m) Front Left side 0.0569 0.0407 0.0569 0.0407
Max. Road holding (m) Front Right side 0.0573 0.0448 0.0573 0.0448
Max. Road holding (m) Rear Left side 0.0551 0.0402 0.0551 0.0402
Max. Road holding (m) Rear Right side 0.0594 0.0448 0.0594 0.0448
Max. seat jerk (m/s
3
) 13.9876 1.5729 13.9846 1.5781
Max. sprung mass jerk (m/s
3
) 12.4447 1.3415 12.4450 1.3448
Natural frequency seat (Hz) 1.2015 1.0902 1.2017 1.0898
Time (sec) 1790 1650 2050 1890
0 1 2 3 4 5 6 7 8 9 10
-0.04
-0.02
0
0.02
0.04
0.06
0.08
Time (sec)
S
e
a
t
D
i
s
p
l
a
c
e
m
e
n
t
( m
)
Passive
Active
Figure 3. Seat displacement v/s time.
0 1 2 3 4 5 6 7 8 9 10
-0.02
-0.01
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
Time (sec)
S
p
r
u
n
g
M
a
s
s
V
e
r
t
i
c
a
l
D
i
s
p
l
a
c
e
m
e
n
t
( m
)
Passive
Active
Figure 4. Sprung mass vertical displacement v/s time.
Optimal Design of Passenger Car Suspension for ...
J. of the Braz. Soc. of Mech. Sci. & Eng. Copyright 2008 by ABCM January-March 2008, Vol. XXX, No. 1 / 73
0 1 2 3 4 5 6 7 8 9 10
-1.5
-1
-0.5
0
0.5
1
1.5
Time (sec)
S
p
r
u
n
g
M
a
s
s
P
i
t
c
h
D
i
s
p
l
a
c
e
m
e
n
t
( d
e
g
r
e
e
)
Passive
Active
Figure 5. Sprung mass pitch displacement v/s time.
0 1 2 3 4 5 6 7 8 9 10
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
Time (sec)
S
p
r
u
n
g
M
a
s
s
R
o
l
l
D
i
s
p
l
a
c
e
m
e
n
t
( d
e
g
r
e
e
)
Passive
Active
Figure 6. Sprung mass roll displacement v/s time.
0 1 2 3 4 5 6 7 8 9 10
-2.5
-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
Time (sec)
S
e
a
t
A
c
c
l
e
r
a
t
i
o
n
( m
/
s
2
)
Passive
Active
Figure 7. Seat acceleration v/s time.
0 1 2 3 4 5 6 7 8 9 10
-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
Time (sec)
S
p
r
u
n
g
M
a
s
s
V
e
r
t
i
c
a
l
A
c
c
l
e
r
a
t
i
o
n
( m
/
s
2
)
Passive
Active
Figure 8. Sprung mass vertical acceleration v/s time.
0 1 2 3 4 5 6 7 8 9 10
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
Time (sec)
S
p
r
u
n
g
M
a
s
s
R
o
l
l
A
c
c
l
e
r
a
t
i
o
n
(r
a
d
/
s
2
)
Passive
Active
Figure 9. Sprung mass roll acceleration v/s time.
0 1 2 3 4 5 6 7 8 9 10
-1.5
-1
-0.5
0
0.5
1
1.5
Time (sec)
S
p
r
u
n
g
M
a
s
s
P
i
t
c
h
A
c
c
l
e
r
a
t
i
o
n
(r
a
d
/
s
2
)
Passive
Active
Figure. 10 Sprung mass pitch acceleration v/s time.
Anil Shirahatt et al
74 / Vol. XXX, No. 1, January-March 2008 ABCM
0 1 2 3 4 5 6 7 8 9 10
-0.04
-0.03
-0.02
-0.01
0
0.01
0.02
0.03
0.04
Time (sec)
S
u
s
p
e
n
s
i
o
n
T
r
a
v
e
l
( m
)
Passive - Front Left
Active - Front Left
Figure 11. Front left suspension travel v/s time.
0 1 2 3 4 5 6 7 8 9 10
-0.035
-0.03
-0.025
-0.02
-0.015
-0.01
-0.005
0
0.005
0.01
0.015
Time (sec)
S
u
s
p
e
n
s
i
o
n
T
r
a
v
e
l
( m
)
Passive - Rear Left
Active - Rear Left
Figure 12. Rear left suspension travel v/s time.
0 1 2 3 4 5 6 7 8 9 10
-0.03
-0.02
-0.01
0
0.01
0.02
0.03
Time (sec)
S
u
s
p
e
n
s
i
o
n
T
r
a
v
e
l
( m
)
Passive - Front Right
Active - Front Right
Figure 13. Front right suspension travel v/s time.
0 1 2 3 4 5 6 7 8 9 10
-0.03
-0.025
-0.02
-0.015
-0.01
-0.005
0
0.005
0.01
0.015
Time (sec)
S
u
s
p
e
n
s
i
o
n
T
r
a
v
e
l
( m
)
Passive - Rear Right
Active - Rear Right
Figure 14. Rear right suspension travel v/s time.
0 1 2 3 4 5 6 7 8 9 10
-0.01
0
0.01
0.02
0.03
0.04
0.05
0.06
Time (sec)
T
y
r
e
D
i
s
p
l
a
c
e
m
e
n
t
( m
)
Passive - Front Left
Active - Front Left
Figure 15. Front left tyre displacement v/s time.
0 1 2 3 4 5 6 7 8 9 10
-0.01
0
0.01
0.02
0.03
0.04
0.05
0.06
Time (sec)
T
y
r
e
D
i
s
p
l
a
c
e
m
e
n
t
( m
)
Passive - Rear Left
Active - Rear Left
Figure 16. Rear left tyre displacement v/s time.
Optimal Design of Passenger Car Suspension for ...
J. of the Braz. Soc. of Mech. Sci. & Eng. Copyright 2008 by ABCM January-March 2008, Vol. XXX, No. 1 / 75
0 1 2 3 4 5 6 7 8 9 10
-0.01
0
0.01
0.02
0.03
0.04
0.05
0.06
Time (sec)
T
y
r
e
D
i
s
p
l
a
c
e
m
e
n
t
( m
)
Passive - Front Right
Active - Front Right
Figure 17. Front right tyre displacement v/s time.
0 1 2 3 4 5 6 7 8 9 10
-0.01
0
0.01
0.02
0.03
0.04
0.05
0.06
Time (sec)
T
y
r
e
D
i
s
p
l
a
c
e
m
e
n
t
( m
)
Passive - Rear Right
Active - Rear Right
Figure 18. Rear right tyre displacement v/s time.
0 1 2 3 4 5 6 7 8 9 10
-15
-10
-5
0
5
10
15
Time (sec)
S
e
a
t
J
e
r
k
( m
/
s
3
)
Passive
Active
Figure 19. Set jerk v/s time.
0 1 2 3 4 5 6 7 8 9 10
-15
-10
-5
0
5
10
15
Time (sec)
S
p
r
u
n
g
M
a
s
s
J
e
r
k
( m
/
s
3
)
Passive
Active
Figure 20. Sprung mass jerk v/s time.
0 1 2 3 4 5 6 7 8 9 10
-15
-10
-5
0
5
10
15
Time (sec)
A
c
t
u
a
t
o
r
F
o
r
c
e
( N
)
Front Left
Figure 21. Front left actuator force v/s time.
0 1 2 3 4 5 6 7 8 9 10
-15
-10
-5
0
5
10
15
Time (sec)
A
c
t
u
a
t
o
r
F
o
r
c
e
( N
)
Rear Left
Figure 22. Rear left actuator force v/s time.
Anil Shirahatt et al
76 / Vol. XXX, No. 1, January-March 2008 ABCM
0 1 2 3 4 5 6 7 8 9 10
-15
-10
-5
0
5
10
15
Time (sec)
A
c
t
u
a
t
o
r
F
o
r
c
e
( N
)
Front Right
Figure 23. Front right actuator force v/s time.
0 1 2 3 4 5 6 7 8 9 10
-15
-10
-5
0
5
10
15
Time (sec)
A
c
t
u
a
t
o
r
F
o
r
c
e
( N
)
Rear Right
Figure 24. Rear right actuator force v/s time.
Conclusion
Considering the power and capabilities of GA, the present work
has attempted to design optimal vehicle suspensions using it. Design
objectives such as maximum bouncing acceleration of seat and
sprung mass, root mean square (RMS) weighted acceleration of seat
and sprung mass as per ISO2631 standards, jerk, suspension travel,
road holding and tyre deflection are introduced for accessing
comfortability of the suspension. While the searching space of the
parameters is very large, the solution space is very tight due to the
presence of various constraints. Therefore, the constrained
optimization problem is converted into unconstrained one using
penalty function approach.
In order to verify the validity of the results, the GA results were
compared to those obtained by simulated annealing technique and
found to yields similar performance measures. This validates the
GA results and also demonstrates that there exists other feasible
design, which is able to achieve the same objective.
From the simulation results, it can be observed that the reduction
of the drivers vertical displacement peak is approximately 74.2% in
case of active suspension as compared with passive suspension and
also settling time is reduced from 6 sec to 3.5 sec. Also the vertical
weighted RMS acceleration of seat and sprung mass is reduced from
0.3032 m/s
2
to 0.0534 m/s
2
and 0.2834 m/s
2
to 0.0492 m/s
2
using
active LQR controller design since more weightage is given ride
comfort. In case of active suspension travel increases by 56-60%
than passive suspension to provide more ride comfort i.e. less
displacement of sprung mass while tyre displacement is reduced by
28.5% to give better road holding, indicating active suspension
system has better potential to improve both comfort and road
holding.
References
Ahmadian, M.T., Sedeh, R.S., and Abdollahpour, R., Application of
Car Active Suspension in Vertical Acceleration Reduction of Vehicle Due to
Road Excitation and Its Effect on Human Health, International Journal of
Scientific Research (In press).
Alkhatib, R., Jazar, G.N., and Golnaraghi, M.F., 2004, Optimal Design
of Passive Linear Suspension Using Genetic Algorithm, Journal of Sound
and Vibration, Vol. 275, pp. 665-691.
Baumal, A.E., McPhee, J.J., and Calamai, P.H., 1998, Application of
Genetic Algorithms to the Design Optimization of an Active Vehicle
Suspension System, Computer Methods in Applied Mechanics and
Engineering, Vol. 163, pp. 87-94.
Bourmistrova A, Storey I and Subic A, 2005, Multiobjective
Optimisation of Active and Semi-Active Suspension Systems with
Application of Evolutionary Algorithm, International Conference on
Modeling and Simulation, Melbourne, 12-15 December 2005.
Gao Huijun, Lam James and Wang Changhong, 2006, Multi-Objective
Control of Vehicle Active Suspension Systems via Load-Dependent
Controllers, Journal of Sound and Vibration, Vol. 290, pp. 654-675.
Gillespie Thomas D., 2003, Fundamentals of Vehicle Dynamics,
Society of Automotive Engineers, Warrendale.
Gobbi, M. and Mastinu, G., 2001, Analytical Description and
Optimization of the Dynamic Behaviour of Passively Suspended Road
Vehicles, Journal of Sound and Vibration, Vol. 245, No. 3, pp. 457-481.
Griffin, M.J., 2003, Handbook of human vibration, Academic press,
New York.
Hemiter Marc E., 2001, Programming in Matlab, Thomson Learning,
Singapore.
ISO: 2631-1, 1997, Mechanical vibration and shock - Evaluation of
human exposure to whole-body vibration.
Mantaras Daniel A., and Luque Pablo, 2006, Ride Comfort
Performance of Different Active Suspension Systems, International Journal
of Vehicle Design, Vol. 40, No. 1/2/3, pp. 106-125.
Ogata, K., 1996, Modern Control Engineering, Prentice-Hall, New
Delhi, 3rd edition.
Panzade, P.K., 2005, Modeling and Analysis of full vehicle for ride and
handling, M.E. Thesis, PSG College of Technology, Coimbatore.
Rettig Uwe, and Stryk Oskar von, 2005, Optimal and Robust Damping
Control for Semi-Active Vehicle Suspension, Proc. ENOC-2005,
Eindhoven, Netherlands, 7-12 August 2005.
Rill, Georg. 2006, Vehicle modeling by subsystems, J. Braz. Soc.
Mech. Sci.& Eng., Vol.28, no.4, p.430-442.
Roumy Jean Gabriel, Boulet Benoit, and Dionne Dany, 2004, Active
Control of Vibrations Transmitted Through a Car Suspension, International
Journal of Vehicle Autonomous Systems, Vol. 2, No.3/4, pp. 236-254.
Sharkawy, A.B., 2005, Fuzzy and Adaptive Fuzzy Control for the
Automobiles Active Suspension System, Vehicle System Dynamics, Vol.
43, No. 11, pp. 795-806.
Sun Lu, 2002, Optimum Design of Road-Friendly Vehicle Suspension
Systems Subjected to Rough Pavement Surfaces, Journal of Applied
Mathematical Modelling, Vol. 26, pp. 635-652.
Wong, J.Y., 1998, Theory of Ground Vehicles, John Wiley and Sons
Inc., New York.
Zaremba, A., Hampo, R., and Hrovat D., 1997, Optimal Active
Suspension Design Using Constrained Optimization, Journal of Sound and
Vibration, Vol. 207, No. 3, pp. 351-364.