Professional Documents
Culture Documents
Electromechanical Active Body Control
Electromechanical Active Body Control
ABSTRACT
We present the systematic development of the module based on specification requirements via
different approaches and concepts. Validation of the simulation results was available after the
first partial and module tests and simultaneously served for confirmation of the specification
requirements. Measurements on the test rig focused in particular on characteristics, travel and
actuating forces, as well as dynamic behavior.
Furthermore, this paper shows the integration and the design of the active electromechanical
system in a demonstrator vehicle. A description of the control algorithm for improving
driving comfort and driving dynamics is given. Driving tests show that the average energy
consumption is below that of a comparable vehicle with an active hydraulic suspension
system. Less energy consumption leads to less fuel consumption and therefore less CO2
emission.
INTRODUCTION
In recent years, electronic suspension control systems have been developed and successfully
introduced into the market. Their purpose is to improve ride comfort as well as lateral
dynamics of the vehicle. These active body and roll control systems are nearly exclusively
based on hydraulic actuators. Their function is well accepted and satisfying, yet hydraulic
systems suffer from various disadvantages, such as high energy consumption, problems with
dirt contamination, and costly assembly of the various components of the system [1].
Furthermore, their compatibility with future hybrid vehicles is limited. The goal of
development was therefore to design and build an active body control system based on
electromechanical actuators.
The system development was a cooperation between ZF and the Research Center of
Volkswagen. ZF was responsible for development and testing of the actuators, including
power electronics and actuator control software, and Volkswagen was responsible for
integration into a demonstrator vehicle, design and application of the control algorithm for
vehicle dynamics, and driving tests.
ACTUATOR DEVELOPMENT
For the desired functionality to be achieved, the vehicle has to be equipped with individually
controllable actuators on each wheel for modulation of the vertical forces between the body
and the tire-contact centers. In contrast to common leveling systems, displacement of the
spring must happen within milliseconds - in compression and in rebound direction. The
required adjustment forces and travels and the available installation space were derived from
an upper-class sedan chosen as demonstrator vehicle. A particular challenge is presented by
the need to integrate the entire drive unit into the suspension dome on both axles with all their
different restrictions.
mbody
Fres
E
M
c2 Fstat
c1 Fact
0
mwheel t
Figure 1 shows schematically the basic configuration of the actuator, which consists of two
series-connected coil springs, a spindle drive, and a parallel hydraulic damper.
Similar to common active hydraulic systems, the spring seat of the suspension spring c1 can
be displaced, which in this case is done by an electric motor. By means of the connected
spindle drive, it is possible to provide both, compressive as well as tensile forces.
To reduce electrical power requirement, a second so-called accumulator spring, which carries
the static vertical load, is added in parallel to the electric motor.
Apart from covering losses, the average power demand is zero, the second advantage over
conventional systems. Only changes in the stationary load conditions of the vehicle result in
permanent setting forces. As these forces are small, the energy needed can constantly be
supplied by the vehicle electrical system.
By these measures, and with the energy recovery capability of the electric motor running
alternately in generator mode taken into account, a substantial reduction of power
consumption is obtainable compared to hydraulic systems, which need at least one
permanently driven pump.
Conceptual design
Figure 2 shows the method applied for approval and evaluation of the optimal conceptual
design in a very early stage of the developing process when a detailed description of the
necessary components was not available.
KK
KK
F2F2StStRoRo
F2F2StSt RoRo
FTFT SpSpSp
FTFT Sp
F1F1 DD
F1F1 DD
1. Collect basic arrangement possibilities, e.g. in parallel, coaxial, or angular to the strut.
2. Arrange main components relative to each other, consider boundary conditions and
desired main functions, e.g. fastening of electric motor to vehicle body, internal or
external rotor, driven part of spindle (nut or screw), and separate load paths.
3. Select optimal conceptual design by means of predetermined assessment criteria, e.g.
low penetration to ensure mechanical integrity, minimal rotor / spindle inertia, static
seals, suitable heat dissipation.
The main functions and basic components are connected in a very abstract mode by working
principles and interactions only, e.g. magnetic force, mechanical force, relative lateral or
rotational degree of freedom, as shown in the left part of Figure 2.
The structure thus created can be varied in many ways, and in a succeeding step it can be
transformed into concrete part shapes and dimensions comparable to other layouts, as shown
in the right part of Figure 2.
Detailed design
Based on the conceptual design mentioned above, the detailed design starts with
dimensioning the main components, i.e. electric motor, coil springs, spindle drive, damper,
stops, and the top mount.
A compromise has to be found to achieve the required dynamic behavior and to fulfil the
limited package specification at the same time. The leading parameters, motor torque and
spindle ratio, have to be considered, as well as limited electrical peak current and minimal
rotor inertia.
Modified top mount
Figure 3 shows a cutaway of the
detailed strut assembly, including the Piston rod
electromechanical actuator. The length
of the strut is determined by the two Electric motor
coil springs, the suspension travel and Accumulator spring
the actuator travel, which in turn
depends on the stiffness of the two Spindle drive
springs. The outer diameter is given by
the required motor torque, spring Rubber bellows
stiffness and actuator travel of the
Displaceable spring seat
accumulator spring. The whole
drivetrain of the actuator is identical on
Compression stop
both axles to enable cost-effective
manufacture.
Suspension spring
As known from passive suspensions,
the damper, the springs and parts of the Shock absorber
housing and strut mounts are adapted to with rebound stop
the different packages, wheel loads, and
suspension ratios. Fixed spring mount
Actuator simulation
The model shown in Figure 4 consists of the electric motor with its characteristics, including
thermal and magnetic losses, the two springs, being displaced by the spindle drive,
considering mass inertia and speed-dependent friction, and the conventional damper strut with
non-linear compression and rebound stops.
motor speed [rad/s]
motor_temp [°C] 5
3 ambient_temp [°C]
rotor speed [rad/s] 2
motor torque [Nm] motor torque [Nm]
1 desired motor torque [Nm]
delta_z strut [m] force spring seat [N] spindel force [N]
spindel position [m] spindle position [m]
suspension strut
ACTUATOR TESTING
A demonstrator vehicle is equipped with four actuators, several sensors, a double-layer high-
capacity condenser and dSPACE hardware for processing the control algorithms and data
recording. There are no roll bars in the demonstrator vehicle.
zspist / zspsoll, -
1
0.8
0.6
0.4
0 1 2 3 4 5
0
Measurement
phase, grad
Simulation
-50
-100
0 1 2 3 4 5
f, Hz
Figure 6: Desired and realized dynamic behavior (PT2 Eigenfrequency = 5.8 Damping ratio = 0.9)
Figure 7 shows an illustration of the system in the demonstration vehicle. Two dSPACE
Autoboxes (3) are used for controlling. One box controls the four electric motors and the
other box processes the vehicle control. The boxes communicate via CAN. The electric motor
control is managed by ZF, while VW is responsible for the vehicle control. The power
electronics (1) are located as close as possible to the actuators in order to minimize the power
loss in the AC power cables. The DC power cables of the electronics are connected to a
double-layer high-capacity condenser (2) located in the middle of the car. The
electromechanical system works with a 12 V on-board net. The condenser has a capacity of
100 F and is used as a storage device in order to stabilize the voltage. It is connected parallel
to the electrical system of the vehicle. The data cables of the electronics are connected to the
ZF Autobox. The VW Autobox is connected to the CAN Bus of the vehicle and several
sensors not shown in Figure 7.
1 1
3
AutoBox
2
3
Electric system
AutoBox
1
Data cable
Power cable
The desired spindle position is a sum of two parts. One part is calculated in one horizontal
controller for all struts and the other part is calculated in a strut individual vertical controller.
CAN-Bus-Data
horizontal controller
sensor signals
The vertical controller computes the desired spindle position caused by road-induced
vibrations of the body. The main parts of the vertical controller are a skyhook algorithm and
an algorithm to tune the vertical eigenfrequency and damping ratio of the body. The needed
sensor signals are the body acceleration at the strut domes and the suspension deflection.
TEST RESULTS
System performance and energy consumption are shown in the next sections.
Vertical dynamics
Figure 9 shows the power spectral density of the vertical accelerations at the suspension-strut
domes. The standard data were measured with the same vehicle equipped with a conventional
suspension system. The vertical accelerations in the frequency range up to 3 Hz of the active
system are significantly smaller than the accelerations of the standard vehicle. In the
frequency range above 3 Hz, the vertical accelerations are nearly the same as in the standard
vehicle. As a passenger you notice the flat ride of the vehicle body like in a sporty car but
without the disturbing higher frequency accelerations.
dome fl dome rl
2.5 2.5
standard
2 2 active
psd (az)
psd (az)
1.5 1.5
1 1
0.5 0.5
0 0
0 5 10 15 20 0 5 10 15 20
f, hz f, hz
Figure 9: Power spectral density of vertical acceleration at the suspension-strut dome front left and rear left,
80 km/h, rough proving ground road
Lateral dynamics
The roll angle of the active system during cornering is significantly smaller than the roll angle
of the standard vehicle with a conventional suspension. The left part of Figure 10 shows the
roll angle. The right side shows the influence of roll moment distribution on the steering
tendency.
Depending on roll moment distribution, different steering angles are needed to achieve a
given lateral acceleration above 4 m/s2. There is no influence below 4 m/s2. Dynamic roll
moment distribution is integrated into the horizontal controller. This leads to a neutral steering
tendency with enough stability reserve during dynamic maneuvers [2].
5 140
standard 90 % Front Axle
active 120
4 30 % Front Axle
100 50 % Front Axle
steering angle, °
roll angle, °
3 80
2 60
40
1
20
0 0
0 2 4 6 8 0 2 4 6 8
2 2
lateral acceleration, m/s lateral acceleration, m/s
Figure 10: Roll angle and steering angle vs. lateral acceleration
The flat ride on rough roads without disturbing higher frequency acceleration, the smaller roll
angle and the neutral steering tendency lead to excellent driving comfort and excellent driving
dynamics.
Energy consumption
Today, several upper-class cars are equipped with active suspension systems. These systems
frequently use hydraulic actuators. The power loss of hydraulic systems is generally higher
than that of electrical systems. Electrical systems have the potential of energy regeneration
and do not require external power if they are not in use. Figure 11 shows data of a test on a
very rough proving ground track. This is the worst case test driven with the demonstrator
vehicle.
steering angle, °
200
0
-200
0 10 20 30 40 50 60 70 80 90 100
10
ay, m/s2
-10
0 10 20 30 40 50 60 70 80 90 100
10
az, m/s2
-10
0 10 20 30 40 50 60 70 80 90 100
motor current, A
400
200
0
vehicle system current, A
-200
0 10 20 30 40 50 60 70 80 90 100
100
50
0
0 10 20 30 40 50 60 70 80 90 100
time, s
Figure 11: Very rough proving ground track with curves (worst case)
The track consists of several curves with high lateral acceleration and a very rough surface
with low and high frequency bumps. Figure 10 shows the high lateral accelerations ay and
vertical accelerations az during the test. The motor current, shown in Figure 11, is the sum of
the DC current of the four motors. It is the output current of the condenser. Positive current
means consumption and negative current means regeneration. The vehicle system current is
the input current of the condenser which is supplied by the electrical vehicle system. The
condenser filters the current peaks significantly. Peaks in the motor current of maximally 300
amperes leads to peaks in the electrical system current of maximally 90 amperes.
The mean current consumption is 36 amperes during this track. This leads to a fuel
consumption of approx. 0.6 l/100 km. In Figure 12, the mean current consumption and fuel
consumption are shown for different road types. The proving ground tracks were driven at the
limit range of vehicle dynamics. The tests on the rough country road and on the highway are
good examples of normal driving situations. In these cases, fuel consumption of the system is
less than 0.2 l / 100 km.
Course description Mean current Fuel consumption
(A) (l /100km)
Very rough proving ground track with curves 36 0.6
Rough proving ground handling course 28 0.5
Rough country road with curves 8 0.2
Highway (160 km/h) 2.5 0.15
Figure 12: Energy consumption of electromechanical active body control system on different road types
In comparison to that, Figure 13 shows the test results with a similar demonstrator vehicle
equipped with a hydraulic active body control system.
CONCLUSIONS / SUMMARY
Tests with the demonstrator vehicle equipped with the electromechanical active body control
show that both driving comfort and driving dynamics can be improved. Pitch and roll induced
by braking and steering are largely avoided. Vertical vibration of the body due to rough roads
is damped in the best way. Vehicle dynamics is optimized by dynamic distribution of the roll
moment. The average energy consumption is significantly below that of a comparable vehicle
with hydraulic active suspension system. Less energy consumption leads to less fuel
consumption and therefore less CO2.
Future development work will have to address the question of packaging for modular axles.
For commercial reasons, a concept with shared components for front and rear axle is more
favorable from the viewpoint of a supplier. Its realization requires close coordination between
supplier and car manufacturer in an early stage of axle development.
Furthermore, future work has to be done on advanced vehicle electrical systems which are
able to supply electrical active suspension systems. In the demonstrator vehicle a double-layer
high-capacity condenser is used to stabilize the voltage. Below the worst-case average
consumption of 36 amperes there are current peaks of nearly 300 amperes which are filtered
significantly by the condenser. This works well in the demonstrator vehicle but is not a
solution for series cars.
REFERENCES
[3] H. Duda, S. Berkner, Integrated Chassis Control Using Active Suspension and
Braking, Avec ’04, Page 347-352