Fu&Cebon 2002
Fu&Cebon 2002
Fu&Cebon 2002
4, 2002 281
E-mail: dc@eng.cam.ac.uk
Reference to this paper should be made as follows: Fu, T.-T. and Cebon, D.
(2002) Analysis of a truck suspension database, Heavy Vehicle Systems, A
Series of the Int. J. of Vehicle Design, Vol. 9, No. 4, pp. 281297.
1 Introduction
In the UK, 80% of goods are transported by road (Dept. of Transport, 1992), mainly by
heavy vehicles. In terms of freight distance travelled (tonne-km), road transport accounts
for 66% of all transportation modes, including air, sea, rail, etc. Heavy vehicles are an
important component of the national economy, but are also responsible for a significant
part of the damage to roads and bridges (Cebon, 1999). Suspension design is a key factor
in the dynamic tyre forces which are thought to contribute to this damage (Cebon, 1999;
Mitchell, 1991; Cole and Cebon, 1996; Woodroofe, 1996), as well as ride comfort
Woodroofe, 1996), roll stability (Winkler et al., 2000), handling (Anon., 1989), and
braking (Hardy and Cebon, 1993).
There is a fundamental trade-off between the ride vibration and roll performance of
heavy vehicle suspensions. In general, in order to increase roll stability it is necessary to
increase vertical stiffness (Cole and Cebon, 1996). The use of anti-roll bars alleviates this
trade-off to some extent, since they can allow the suspension to be soft in the vertical
direction, but stiff in roll (Cole and Cebon, 1996).
Key factors that influence roll stability are the height of the centre of gravity (CG)
and track width; the compliance (and any back-lash) of the suspensions and tyres;
movement of the payload (e.g. sloshing of fluid in a tanker); and the torsional flexibility
of the vehicle frame (see review by Winkler (2000)).
Ride and dynamic tyre forces (road damage) are strongly influenced by the stiffness
and damping of the suspension (Cole et al., 1996). Softer suspensions generally have
lower natural frequencies and generate lower dynamic tyre forces than stiffer
suspensions (Cebon, 1999; Gyenes et al., 1992).
In this paper, measured suspension data from the Transportation Research Institute of
the University of Michigan (UMTRI) are analysed (Fancher et al., 1980; Fancher et al.,
1986; Winkler et al., 1992). Analytical models and assumptions are employed to fill the
gaps in the data. Analyses presented here aim to elucidate trends in the stiffness
characteristics, natural frequencies and roll performance achieved by current
suspensions.
Rollover is a specific problem in heavy vehicle operations due to their large payload and
high centre of gravity (CG). The problem can occur while the vehicle is in steady turning
or a transient handling manoeuvre, such as a lane change. The main parameters involved
are the CG height, payload shifting (e.g. liquid sloshing in tanks), suspension and tyre
compliance. The propensity to rollover during steady turning can be characterised by the
static rollover threshold. This is the steady lateral acceleration which causes rollover of
the vehicle (Fancher, 1985).
A simple static roll plane model was used to predict the rollover threshold of vehicles
from measured suspension parameters. As illustrated in Figure 1, the model has two
degrees of freedom: sprung mass and unsprung mass roll angles ( r and t ). The effect
of spring stiffness on roll motion is combined with auxiliary roll stiffness ( K raux ) into
the total roll stiffness ( K rtot ) of the suspension according to the following relation:
m s Ay
ms g
hs
rtot
t
h rc
m u Ay
hu
m ug
Krt Degrees of freedom
where r and t are the roll angles of the sprung and unsprung masses. Arranging these
into a matrix form and using rollover threshold Ay and sprung mass roll angle r as the
state variables gives:
Ay msg(hs hrc )(K rt ms ghrc mughu ) + Krtot (ms ghs + mughu Krt )
= t
g ms g(hs hrc )(ms ghrc + mughu ) Krtot (ms ghs + mughu ) (5)
The rollover threshold is defined as the lateral acceleration when one wheel lifts off
the ground. In this condition, the unsprung mass roll angle is:
(ms + mu )g
t =
2 Kt t (6)
2
where 2t is the track width, and Kt is the tyre stiffness. (Note that Krt = 2Kt t .)
Substituting Equation (6) into Equation (5) gives the rollover threshold.
Note that this simplified model was chosen specifically for comparing the measured
suspension data. It is not dynamically similar to any particular heavy goods vehicle,
because: (i) it only has one axle, and therefore does not account the effects of multiple
axles, including roll/torsion and bounce/pitch coupling; (ii) its inertia is all concentrated
above the suspension. Nevertheless the model is representative of the broad class of
heavy vehicles to which the measured suspensions are normally attached.
No. of suspensions
25 23
20
15 12
11
10
6
4
5
2 2
1
0
drive- drive- drive- drive- steer- steer- Trailer- Trailer-
air steel torsion walking steel walking air steel
50 47
40
30
20
10
10
2 2
0
Europe- Europe- North North
commercial military America- America-
commercial military
Market origin
Table 2
drive single steel flat (9) NA 71.2 10.73 1.16 8.06E+05 780
drive single steel flat (14) NA 93.4 9.69 2.26 1.15E+06 719
drive tandem 4-spring steel flat (7) NA 80.05 6.17 1.13 8.18E+05 677
drive tandem 2-spring steel flat (9) Eur 77.85 9.19 1.24 9.57E+05 657
drive tandem 2-spring steel taper (2) NA 75 10.8 0.8 1.19E+06 863
drive tandem 2-spring steel taper (3) Eur 77.85 7.22 1.36 6.82E+05 334
drive tandem walking-beam flat NA, M 71.15 5.12 2.03 3.69E+05 930
drive tandem torsion-bar torsion-bar NA 71.15 8.11 0.79 8.73E+05 638
steer single steel flat (6) Eur 62.3 2.52 1.24 2.13E+05 515
steer single steel flat (9) NA 53.4 2.77 1.47 2.45E+05 473
Ver. Stiff. : Vertical stiffness of the spring ( k s ).
steer single steel flat (10) NA 64.9 2.88 1.07 3.04E+05 473
R C H : Roll centre height of the suspension ( hrc ).
steer tandem walking-beam flat NA, M 62.3 4.87 1.69 3.79E+05 927
trailer single trailing-arm air overslung NA 80.1 12.32 10.73 4.06E+05 557
Tot. R. Stiff. : Total roll stiffness of the suspension ( K rtot ).
Analysis of a truck suspension database
trailer single trailing-arm air (iso) overslung NA 89 17.74 13.56 3.62E+05 571
Aux. R. Stiff. : Auxiliary roll stiffness of the suspension ( K raux ).
trailer single trailing-arm air ind Eur 89 14.24 10.17 4.86E+05 100
trailer tandem trailing-arm air underslung NA 71.15 8.93 6.08 3.40E+05 615
vertical spring stiffness and suspension roll centre height.
trailer tandem 4-spring steel flat NA 71.15 13.56 1.13 1.41E+06 795
trailer tandem 4-spring steel flat Eur 84.5 16.16 2.26 1.47E+06 736
Suspension experimental data from Winkler (1992; 1997).
trailer tandem 4-spring steel taper (3) NA 71.15 11.72 2.95 1.05E+06 592
Rated L. : Rated load of the suspension. It is also the testing load applied in the measurements.
axle and spring for the suspensions, their origins and applications, etc. It also includes
from the UMTRI databases. It includes descriptive information which gives the types of
quantitative information: the rated load, total and auxiliary suspension roll stiffness,
287
288 T.-T. Fu and D. Cebon
Table 3 Samples of calculated data for the suspensions listed in Table 2. See Figure 1 for
definition of the variables. Additional parameters are: k t - tyre stiffness, and t - half wheel track.
Analysis of a truck suspension database 289
1 ke
fn =
2 mtesting (7)
where, k e is the effective vertical stiffness for the combination of spring and tyres (N/m)
and mtesting is the equivalent mass (kg) derived from the suspension test load.
Various analyses were performed on the database. Parametric analyses were performed
to explore relationships between various suspension stiffnesses, in order to elucidate their
contributions to the total roll stiffness. The conflict between vehicle ride and roll
performance was also investigated from the suspension design perspective. This conflict
is demonstrated by the interrelationship between natural frequency and rollover
threshold. The third type of analysis examined the joint effect of total roll stiffness and
roll centre height on rollover threshold, in order to understand trends in suspension
design.
1.2E+6 Drive-air
1.0E+6 Drive-steel
8.0E+5 Drive-torsion
6.0E+5 Drive-walking
4.0E+5 Steer-steel
Line of Equity
2.0E+5 Steer-walking
0.0E+0 Trailer-air
0.0E+0 2.0E+5 4.0E+5 6.0E+5 8.0E+5 1.0E+6
Trailer-steel
Aux. roll stiffness (N-m/rad)
80
70
60
50
40
30
20
10
0
Drive Drive Drive Drive Steer Steer Trailer Trailer
-air -steel -torsion -walking -steel -walking -air -steel
Type of suspensions
1.2E+6 Drive-air
1.0E+6 Drive-steel
8.0E+5 Drive-torsion
6.0E+5 Drive-walking
4.0E+5 Steer-steel
2.0E+5 Steer-walking
0.0E+0 Trailer-air
0.0E+0 5.0E+5 1.0E+6 1.5E+6
Trailer-steel
Vertical stiffness (N/m)
Steel steering axle suspensions obtain their auxiliary roll stiffness by mounting the
axle off the centre of the steel leaf span, and by twisting the leaf springs about the
longitudinal axis. Since they have lower static load and more design emphasis on ride
comfort, their vertical spring stiffness tends to be significantly lower than that of drive
suspensions. For these reasons the percentage of auxiliary roll stiffness in the total roll
stiffness of steering axle suspensions is higher than that for other axle types of steel
suspensions. Therefore, as can be seen in Figures 4 and 6, the stiffness characteristics of
steel steering axle suspensions are more like those of air suspensions than the other steel
suspensions.
According to Anon. (1989), steel suspensions normally do not have additional roll
stiffening mechanisms deliberately incorporated in their designs. This is why they
generally have a low proportion of auxiliary roll stiffness. However, the anti-roll bar
effect of rigid axle and trailing arm assemblies provides very stiff roll resistance for air
suspensions. Their auxiliary roll stiffness is not only high in proportion to total roll
stiffness (because their vertical stiffness is normally low), but also high in absolute terms.
0.5g. This would appear to indicate that suspension designers aim for a specific level of
rollover performance.
Ay/g()
0.80 Drive-air
Drive-steel
0.70
Drive-torsion
0.60
Drive-walking
0.50
Steer-steel
0.40
Steer-walking
0.30 Trailer-air
1.00 1.50 2.00 2.50
Trailer-steel
Natural Frequency (Hz)
quantities in the working database. In this section, the relationship between these two
parameters and the calculated rollover threshold of the suspensions is investigated.
Figures 8 and 9 show the roll-centre height plotted against total roll stiffness for the
suspensions in the database, together with contours of constant rollover threshold Ay/g,
plotted using Equations (5) and (6). Figure 8 shows drive and trailer suspensions, while
Figure 9 shows steering axle suspensions. The values of rollover threshold are indicated
by numbers plotted on the contours. The contours span the range of from zero to an
upper limit Ay/g max , which was calculated from Equations (5) and (6), assuming
infinite suspension roll stiffness Krtot . This gives:
Ay Krt (m + mu )g
= 1 s
g max ms ghs + mu ghu 2Kt t
(8)
This limit is the highest rollover threshold that suspensions of a specific axle type can
achieve. For the drive axles and trailer axles in Figure 8, using the parameters of the
baseline vehicle, the maximum rollover threshold is Ay/g max = 0.5574. For the steering
axle suspensions shown in Figure 9, the limit is Ay/g max = 0.8009.
1700
1500
1300
1100
Drive-air
900 Drive-steel
Drive-torsion
700
Drive-walking
500 Trailer-air
Trailer-steel
300
Stability limit,
Ay/g = 0.0 0.3 0.4 0.5 0.55
100
1.00E+05 1.00E+06 1.00E+07 1.00E+08
1200
Upper limit, Ay/g = 0.8009
1000
0.75
0.7
0.6
800
0.5
600
400
200 Steer-steel
Steer-walking
Note that Equation (8) does not contain the roll stiffness K rtot or roll centre height
hrc . So the limits can be determined simply from the relevant geometry, masses and tyre
stiffnesses.
In order to make a comparison between drive and trailer suspensions, the same
baseline axle loads and CG heights were assumed in the calculations of the rollover
threshold for all suspensions.
There is no strong relationship between roll-centre height and total roll stiffness for
the various suspensions in Figure 8. Some correlation exists between these parameters
for steel suspensions, because the roll centre heights are determined by the geometry of
leaf springs. Stiffer multi-leaf steel suspensions normally require more spring leaves,
which pushes the roll-centre to a higher location. On the other hand, because the roll
centres of air suspensions are determined by the geometric configurations of components
which are not normally influenced by the stiffness of the air bags, the location of roll
centre height and the suspension roll stiffness are independent. One particular trailer air
suspension has the lowest roll centre height, at about 100 mm. It is an independent air
suspension for which the roll mechanism is different from the other trailer air
suspensions of the rigid axle type.
The results for steering axle suspensions are plotted separately in Figure 9 because
their loading conditions are quite different from those of the other two axle types. Their
roll centre heights are essentially constant, and independent of roll stiffness. This is
because their values of vertical spring stiffness are very similar to one another, as
demonstrated in Figure 6. All of these suspensions have similar leaf spring
configurations, which keep their roll centre locations at approximately the same level.
296 T.-T. Fu and D. Cebon
The constant roll-centre height is also a consequence of the requirement to minimise roll-
steer effects generated by the steering linkage. As shown in this figure, the roll centre
heights are all about 450 mm. Data points for the tandem walking-beam suspensions lie
outside the range of data for steel suspensions. The walking-beam steering axle
suspensions have an entirely different duty from other suspensions (particularly for
military usage), with a different steering mechanism, and a different mechanism defining
the position of the roll centre. This leads to their significantly different performance, as
shown in Figure 9.
5 Conclusions
1 Analyses in this paper have explored the characteristics of various existing
suspensions. Both experimental data and analytical methods were applied to generate a
complete suspension database. Levels of theoretical performance were investigated for
different suspension designs.
2 The vertical stiffness of an air spring does not contribute much to its total roll
stiffness in a typical air suspension system. Therefore, it has little effect on the rollover
threshold. Steel leaf spring suspensions have more influence on the rollover performance.
3 One of the design trends of a suspension system is to separate the functions of
ride and roll performance into different components. In particular, the duty of spring
elements in providing roll stiffness can be released and the conflict in achieving both
good ride comfort and rollover performance can be partially resolved. Typical examples
are the introduction of anti-roll bars in steel suspensions and the auxiliary roll stiffening
mechanisms in air suspensions.
4 The design of suspensions for steering axles is governed by factors other than
rollover performance, so the rollover performance normally exceeds practical operation
requirements.
5 Designers of drive axle and trailer axle suspensions appear to aim for a rollover
threshold of 0.4 to 0.5g, and achieve this with a wide range of design parameter variants.
Acknowledgements
The authors are very grateful to Mr Chris Winkler of UMTRI for providing valuable
suspension data. Thanks also to the Ministry of Education in Taiwan, ROC for the
financial support to Dr Fu during this project. This work was also supported by the
members of the Cambridge Vehicle Dynamics Consortium. At the time of writing, the
Consortium consists of the Universities of Cambridge and Cranfield together with
industrial partners from the European vehicle industry: Tinsley Bridge Ltd,
ArvinMeritor, Koni BV, Qinetiq, Pirelli, Shell, Volvo Global Trucks, General Trailers,
Mektronika Systems and Fluid Power Design.
References