August 2020, Vol. 20, No.

3 MANUFACTURING TECHNOLOGY ISSN 1213–2489

DOI: 10.21062/mft.2020.052 © 2020 Manufacturing Technology. All rights reserved. http://www.journalmt.com

Dynamic Simulation Analysis of a Motorcycle Suspension System – Assessment

of Comfort

Stefan Segla, Sayantan Roy

Faculty of Mechanical Engineering, Technical University of Kosice. Letna 9, 042 00 Kosice. Slovakia. E-mail:
[email protected]

The paper deals with dynamic in-plane simulation analysis of a motorcycle suspension. The motorcycle᾿s
mechanical model is considered as a visco-elastically suspended rigid body. Two types of the kinematic
excitation are considered ‒ a deterministic „hat“ shaped bump and stochastically uneven road characte-
rized by its power spectral density. The simulation results for both the deterministic bump and stochas-
tically uneven road show that significant reduction of the root mean square value of the motorcycle body
centroid acceleration (comfort criterion) can be achieved by placing the lower end point of the rare
spring-damper module closer to the beginning of the swinging arm and also by increasing deviation (tilt)
of the spring-damper module from the vertical. The maximum improvement in the root mean square
value of the motorcycle body centroid acceleration is 51.7 % for the deterministic „hat“ shaped bump and
37.4 % for the stochastically uneven road. The method presented in the paper can be employed in design
of both touring motorcycles, which are characterized by higher requirements of comfort, and off-road
motorcycles where protection from impacts generated by bumps is important.

Keywords: Motorcycle, Suspension System, Simulation, Road Profile, Comfort

shock absorber is proposed in [1]. It consists of a non-

Introduction linear parametric model and a black-box neural-
A motorcycle can be considered as a rigid body network-based model. The shock absorber model is
connected to the front and rear wheels with spring- experimentally validated. In [2] simulation and experi-
damper modules. The rigid motorcycle body, compo- mental investigation of motorcycle dynamic characte-
sed of chassis, engine, steering head and rider, consti- ristics for displacement-sensitive shock absorber is
tutes the sprung mass. The wheels and masses atta- presented. Dynamics behaviours of motorcycles in
ched to them constitute the unsprung masses. A mo- turns is studied in [3] and [4]. Sharma and Limbeer in
torcycle suspension serves the following main purpo- [5] and [6] describe a design methodology for the su-
ses: spension of a novel aerodynamically efficient mo-
 provides comfort to the rider by keeping him torcycle. Optimization of passive vehicle suspensions
comfortably isolated from the vibrations ge- for improving the ride comfort of a motorcycle driver
is investigated in [7]. The paper [8] deals with mo-
nerated by the interaction of the wheels with
delling, dynamic analysis and optimization of stiffness
the uneven road, and damping parameters of the main suspension of
 ensures wheel grip on uneven road in order to the Škoda 21Tr trolleybus. In [9] it is shown how the
transmit the required driving, braking and kinematic excitation caused by uneven roads together
also lateral forces. with the design parameters of a vehicle suspension
(stiffnesses of the suspension springs, damping coeffi-
cients and tire stiffnesses) affect the comfort of the
The degree of required comfort depends on the
driver and passengers, safety of the ride and relative
use of the motorcycle. Comfort of racing motorcycles
displacements between the sprung and unsprung mas-
is less important than the motorcycle᾿s capability of ses. In [10] modelling and optimization of a passive
keeping the wheels in the sufficient contact with the seat suspension with a vibration absorber is investiga-
road. Touring motorcycles are characterized by higher ted.
requirements of comfort. In off-road motorcycles the
suspension serves to isolate the sprung mass from im- Mechanical and Mathematical Model
pacts generated by jumps. Suspensions in these mo-
torcycles have greater wheel travel than in touring mo- The motorcycle᾿s mechanical model, considered as
torcycles. a viscoelastically suspended rigid body [11], [12] con-
A grey-box model of a motorcycle mono-tube sists of the motorcycle rigid body (sprung mass) 1, the

indexed on: http://www.scopus.com 373

August 2020, Vol. 20, No. 3 MANUFACTURING TECHNOLOGY ISSN 1213–2489

front (2) and rare (3) wheels (unsprung masses) and the center of the wheels), z3 is the vertical displace-
the front and rare spring-damper modules (4, 5)
ment of the sprung mass centroid T3 and  3 defines
between the sprung and unsprung masses, Fig. 1.
The motorcycle motion can be described by four the pitching angular displacement of the sprung mass.
The equations of motion of the motorcycle can be
independent coordinates z1 , z2 , z3 ,  3, Fig. 2, where
described by the following set of ordinary differential
z1 and z2 are the vertical displacements of the front equations
and rare unsprung masses (the equivalent masses at
z1  b1,eq ( z3  l13  z1 )  k 1,eq ( z3  l13  z1 )  k 3 ( z1  u1 )  0,
m1 (1)

z2  b 2,eq ( z3  l2 3  z2 )  k 2,eq ( z3  l2 3  z2 )  k 4 ( z2  u2 )  0,

m2  (2)

z3  b1,eq ( z3  l1 3  z1 )  b 2,eq ( z3  l2 3  z2 )  k 1,eq ( z3  l1 3  z1 )  k 2,eq ( z3  l2 3  z2 )  0,
m3  (3)

I33  b1,eq l1 ( z3  l13  z1 )  b2,eq l2 ( z3  l2 3  z2 )  k1,eq l1 ( z3  l1 3  z1 )  k 2,eq l2 ( z3  l2 3  z2 )  0, (4)

where the coordinates u 1 and u 2 describe the

uneven road profile (kinematic excitation), Fig. 2.

Fig. 2 Mechanical Model of the Motorcycle

2.1 Equivalent Stiffness and Damping Coefficients -
Front Suspension
Fig. 1 Motorcycle Components
The potential energy of the original tilted spring k1
, Fig. 3, can be expressed by the approximate equation
1 1
 
2 2
V1,or  k1  z3 l1 3  cos 1 z1 cos 1   k1 z3  l1 3  z1 cos2  1 (5)
2 2
and the potential energy of the equivalent spring
k 1, e q is

1
 
2
V1,eq  k1,eq z3  l1 3  z1 . (6)
2

Equality of the potential energy V1,o r and V1,eq le-

ads to the dependence between the stiffness coeffi-
cient k 1, e q of the equivalent spring and the stiffness
coefficient of the original spring k 1

k1,eq  k1 cos2 1. (7)

Using the dissipative function it is possible to de-

rive analogical equation for the damping coefficient
Fig. 3 Determining the Equivalent Stiffness and Damping
b1,eq  b1 cos2 1. (8) Coefficients – the Front Suspension

374 indexed on: http://www.scopus.com

August 2020, Vol. 20, No. 3 MANUFACTURING TECHNOLOGY ISSN 1213–2489

2.2 Equivalent Stiffness and Damping Coefficients - k 2 , Fig. 4, can be expressed by the approximate equa-
Rare Ssuspension tion
The potential energy of the original tilted spring
2
1 

1

 2
V2,or  k2  z3 l2  3  cos 2  x cos 2   k2 z3  l2 3  x cos2  2 . (9)
2    2

The potential energy of the vertical spring k2,eq is

1 2
V2,eq  k2,eq  z3 l2  3  x  . (10)
2
Equality of V 2,or and V2,eq gives

k2,eq  k2 cos2  2 . (11)

Fig. 5 Determining the Equivalent Stiffness and Damping
Using dissipative function we get Coefficients of the Rare Suspension, part B

For the potential energy of the spring k 2 ,eq of the

equivalent case, Fig. 5, the following equation holds
1
V2,eq  k2,eq z 22. (16)
2
Equality of equations (16) and (15) gives
2
 L1 
k2,eq  k2,eq   . (17)
 L

Substituting k2,eq , equation (11), into equation

Fig. 4 Determining the Equivalent Stiffness and Damping (17) gives
Coefficients of the Rare Suspension, part A
2
 L1 
b 2,eq  b2 cos2  2 . (12) k2,eq  k2  2
 cos  2 . (18)
L
 
Now the vertical spring-damper module ( k2,eq , Using dissipative function we get
b 2,eq ) will be replaced by the spring-damper module ( 2
 L1 
k 2,eq , b 2,eq ) acting on the mass m 2 , Fig. 5. b 2, eq  b 2   cos 2  2 . (19)
 L 
For the potential energy in the spring k2,eq we can
write, Fig. 5 Deterministic and Stochastic Road Profi-
les
1
V2,or  k2,eq x 2 , (13) The most important source of excitation is caused
2
by the road unevenness. It can be divided into two ba-
where sic types - deterministic and stochastic ones, as
described e.g. in [13].
L1
x z2 . (14) A suitable deterministic non-normalized road
L unevenness is of the so-called „hat“ shape, which to
Substituting equation (14) into equation (13) we get some extent respect the actual shape of the tyre. It su-
its to the point road-tyre contact modelling. Mathema-
2
1  L1  tical model of the bump, depending on the longitudi-
V2,or  k2,eq z 22   . (15) nal coordinate x , is described by equation
2  L 

indexed on: http://www.scopus.com 375

August 2020, Vol. 20, No. 3 MANUFACTURING TECHNOLOGY ISSN 1213–2489

hm  2 
u x   1  cos d x  for 0  x  d , otherwise u  x   0, (20)
2  

where is: R - radius of the circumscribed circle, the bump. In Matlab simulation [4], the time depen-
Fig. 6, d - the length of the bump, hm - the height of dence of the instantaneous bump height will be used
hm  2 
u t    1  cos T t  for 0  t  T , otherwise u  t   0, (21)
2  

Tab. 1 The rms Values of the Motorcycle Body Centroid Ac-

celeration, Displacement and Maximum Displacement Ampli-
tude - the Ride Over the „Hat“ Shaped Bump
L1 z3 , e f z 3,ef z 3, m ax
2 
 
[m] [m/s2] [m] [m]
0 0.2 0.02615 0.00036 0.025
20 0.2 0.02646 0.00035 0.025
40 0.2 0.02767 0.00035 0.023
0 0.4 0.03754 0.00039 0.040
20 0.4 0.03504 0.00039 0.039
Fig. 6 Deterministic „hat“ shaped bump 40 0.4 0.02856 0.00038 0.038
0 0.6 0.05412 0.00039 0.048
where T  d / v, t  x / v, v is the motorcycle 20 0.6 0.05176 0.00039 0.047
speed and T is the bump transit time. 40 0.6 0.04340 0.00039 0.043
Modelling of the stochastic road is based on the Tab. 2 The rms Values of the Motorcycle Body Centroid Ac-
idea that each stochastic stationary process can be re- celeration and Displacement - the Ride Along a Stochastically
presented by the sum of cosine functions. The theory Uneven Road
of the method [13] was developed in 1971 by z3,rms [m/s2]
 2    L 1 [m] z3,rms [m]
Shinozuka.
0 0.2 0.04625 0.00059
Results of Simulations 20 0.2 0.04598 0.00060
40 0.2 0.04511 0.00063
Parameters of the motorcycle mechanical model 0 0.4 0.05631 0.00056
are: m1  15 kg, m2  18 kg, m3  194 kg, 20 0.4 0.05459 0.00056
I 3  38 kg  m , k1  15000 N/m,
2
k 2  24000 N/m, 40 0.4 0.05022 0.00057
k3  180000 N/m, k 4  180000 N/m, 0 0.6 0.07208 0.00055
l1  0.64 m, 20 0.6 0.06918 0.00055
b 1  7 1 0 N  s/m , b 2  1171 N  s/m , 40 0.6 0.06084 0.00056
l 2  0.7 m,  1  2 7 , H  0.6 m, L  0.6 m.

For the
„hat“ shaped bump the following prameters were Graph 1 shows the dependence of the centroid T3
vertical displacement z 3 on time for the ride over the
used: d  2.68 m, hm  0.06 m, v  11 m/s. When dri-
ving on a stochastically uneven road the parameters as „hat“ shaped bump for two different cases described
follows were used: S u   0   2 2 .1 0  6 m 3 (power in the graph.
The maximum improvement in the rms value of
spectral density of an asphalt-concrete road of average
the motorcycle body centroid acceleration is 51.7 %
quality), v  20 m/s. for the deterministic „hat“ shaped bump and 37.4 %
In Tab. 1 root mean square (rms) values of the mo- for the stochastically uneven road. The rms value of
torcycle body centroid acceleration and displacement the body centroid displacement remains practically
and maximum displacement amplitude for combinati- constant.
ons of different values of 2 parameters  2 and L 1 for
the ride over the „hat“ shaped bump are shown. Conclusions
Tab. 2 shows the rms values of the motorcycle The results presented in the paper show that in
body centroid acceleration and displacement for com- both ride over a „hat“ shaped bump and along a sto-
binations of different values of the same parameters chastically uneven road significant reduction of the
 2 and L 1 as in the previous case, for the ride along a
rms value of the motorcycle body centroid can be
stochastically uneven road. achieved by placing the lower end point of the spring-

376 indexed on: http://www.scopus.com

August 2020, Vol. 20, No. 3 MANUFACTURING TECHNOLOGY ISSN 1213–2489

damper module closer to the beginning of the swin- KARANAM, V.M., CHATTERJEE, A.
ging arm. Significant reduction of the value can be also (2011). Common underlying steering curves for
achieved by increasing deviation (tilt) of the spring- motorcycles in steady turns. In: Vehicle System
damper module from the vertical, but this reduction is Dynamics, Vol. 49, No. 6, pp. 931–948. Taylor &
decreasing with placing the lower end point of the rare Francis. ISSN: 0042-3114.
spring-damper module closer to the beginning of the
Optimization Toolbox For Use With
swinging arm. The rms value of the body centroid dis-
MATLAB. Users Guide (2000), The
placement remains practically constant. The above
MathWorks, Inc.
conclusions are important for improving comfort of
the motorcycle rider, because the centroid acceleration SHAMMA, A., LIMEBEER, D.J.N. (2012).
is a decisive factor determining the motorcycle com- Motorcycle design using matrix inequalities and
fort level. passivity constraints. In: Vehicle System Dynamics,
Vol. 50, No. 3, pp. 377–393. Taylor & Francis.
ISSN: 0042-3114.
SEGLA, S. (2004). Optimization of passive ve-
hicle suspension systems for improving the ride
comfort. In: Acta Mechanica Slovaca, Vol. 8, No.
3C, pp. 161–170. Faculty of Mechanical Engi-
neering - Technical University of Košice. ISSN:
1335-2393.
SEGLA, S., KAMPO, J., SOUKUP, J. (2019).
Dynamic analysis and optimization of the
planar model of the trolleybus Škoda 21Tr. In:
Graph 1 Dependence of on Time - „Hat“ Shaped Bump Manufacturing Technology, Vol. 19, No. 3, pp.
z3
487–491. J. E. Purkyne University in Ústí nad
Acknowledgement Labem. ISSN: 12132489.
The authors were supported by grant project SEGLA, S., KAMPO, J. (2018). The role of
VEGA No. 1/0290/18. modelling of road unevennesses in vehicle dy-
namics. In: Manufacturing Technology, Vol. 18, No.
References 1, pp. 124–129. J. E. Purkyne University in Ústí
BEGHI, A., LIBERATI, M., MEZZALIRA, nad Labem. ISSN: 12132489.
S., PERON, S. (2007). Grey-box modelling of SEGLA, S., OREČNÝ, M., TRISOVIC, N.
a motorcycle shock absorber for virtual proto- (2013). The role of modelling of road uneven-
typing applications. In: Simulation Modeling nesses in vehicle dynamics. In: Manufacturing
Practice and Theory, Vol. 15, pp. 894–907. El- Technology, Vol. 13, No. 4, pp. 530–534. J. E.
sevier B.V. ISSN: 1569-190X. Purkyne University in Ústí nad Labem. ISSN:
LEE, Ch.T., MOON, B.Y. (2006). Simulation 12132489.
and experimental validation of vehicle dynamic JAZAR, R.N. (2008). Vehicle Dynamics: Theory
characteristics for displacement-sensitive shock and Application, p. 1020. Springer, Melbourne.
absorber using fluid-flow modelling. In: Mecha- ISBN: 978-0-387-74243-4.
nical Systems and Signal Processing, Vol. 20, pp.
373–388. Elsevier Ltd. ISSN: 0888-3270. COSSALTER, V. (2002). Motorcycle Dynamics. p.
421. Vittore Cossalter, Padova.
SHARP, R.S. (2012). Rider control of a mo-
torcycle near to its cornering limits. In: Vehicle NIGAM, N.C., NARAYANAN, S. (1994). Ap-
System Dynamics, Vol. 50, No. 8, pp. 1193–1208. plications of Random Vibrations, p. 557. Springer,
Taylor & Francis. ISSN: 0042-3114. New York. ISBN: 978-3540198611.

indexed on: http://www.scopus.com 377