Bilinear Optimal Control of Semi-active Suspension
with ER Damper
1. Introduction
In recent years, active suspension or semi-active suspension have been adopted to vehicles
and buildings to improve ride-comfort and stability. Although active suspension can reduce
vibration more effectively, the equipments cost and energy consumption are not desirable,
and when a control unit breaks, the safety of the system can't be guaranteed. On the other hand,
the semi-active suspension in which the coefficient of viscosity is controlled can avoid those
problems.
In this study, we treated such kinds of semi-active suspension system, and arm to
investigate the possibility of ER fluid, provide an effective control method, and establish a
practical useful control system for vertical and horizontal vibration of vehicles.
This paper is the first step of our study. We introduce a new mechanism which uses ER
fluid in the damper, an effective control method based on bilinear optimal control method, and
the performance tests with experiments. The results suggested the possibility of the method.
2. Semi-active Suspension System with ER Damper
Figure 1 shows the image of experimental system where a balance structure is used to make
experiment easier.
Figure 1: Experimental System
The principle of ER damper is to utilize the characteristic that the viscosity of fluid changes
with voltage charged on. The ER damper cooperating with spring composes the single
semi-active suspension in the set.
It is clear that the kinematics equation of the system can be treated as follows:
m&x& + C ( x& x& ) + k ( x x ) = 0
(1)
where C consists of uncharged part c 0 and charged part c(v ) .
C = c 0 + c (v )
m is mass of M; k is spring constant, x and x are displacements.
3. Control Design
For system (1), we adopt the bilinear optimal theory to design a control rule. As usual doing,
we represent equation (1) as
X(t ) = AX(t ) + BV (t )u (t ) + (t )
(2)
where
x
X(t ) =
x&
0
, u (t ) = c(t ) , V (t ) = x& , A = k
m
1
c
0
m
0
B = x&
m
(t ) = &x&
, (t ) is disturbance from road condition, which is supposed to be a white noise.
Now, we can state out control problem as:
[ find a control u such that
1. Closed loop system is stable.
2. The value of the following criterion function is minimum.
J = [ x T (t )Qx(t ) + u T (t )Ru (t )]dt
(3)
here Q and R are weightings. ]
Omitting the detail derivation, we present an answer u = u to the problem as follows.
u (t ) = V T R 1 B T [P(t )] x(t )
where P is obtained by on-line solution of Riccati differential equation:
T
(4)
T
1
2 x T (t )P(t )V T B T R 1 2 x T (t )P(t )V T B T
4
T
T
T
T
= x (t )Qx(t ) + x (t )P(t )Ax(t ) + x (t )AP(t ) x(t ) x T (t )P(t )VBR 1 B T V T P(t )x(t )
x T (t ) P(t )x(t ) = x T (t )Qx(t ) + 2 x T (t )P(t ) Ax(t )
5)
4. Experimental Results
This section shows some results about basic experiment of ER fluid and performance
tests of the system.
The measured viscosity coefficient of ER can be approximated as
c(v ) = 1.1 10 3 V + 0.5059
(6)
Figure 2: Control Results
Figure 2 is the control results. In those experiments, the weighting parameters are selected
as
30 0
Q=
, R = 1
0 1
The results show that the ER damper has good sensitivity, and semi-active damping system
with ER damper and bilinear controller can reduce vibration more effectively.
5. Conclusion
In this study, we constructed a semi-active suspension system with ER damper and bilinear
optimal controller. The results suggest that such kind of system is expectable. We intend to
develop a practical suspension system for vehicle depends on those results.
