Hindawi Publishing Corporation

Advances in Mechanical Engineering

Volume 2014, Article ID 251594, 13 pages
http://dx.doi.org/10.1155/2014/251594

Research Article
Stability Analysis of a Linear Friction-Induced Vibration Model
and Its Prevention Using Active Force Control

S. M. Hashemi-Dehkordi,1 A. R. Abu-Bakar,2 and M. Mailah1

1
Department of Applied Mechanics, Faculty of Mechanical Engineering, Universiti Teknologi Malaysia,
81310 Johor, Skudai, Malaysia
2
Department of Automotive Engineering, Faculty of Mechanical Engineering, Universiti Teknologi Malaysia,
81310 Johor, Skudai, Malaysia

Correspondence should be addressed to S. M. Hashemi-Dehkordi; s.mahdi.hashemi.d@gmail.com

Received 1 June 2013; Revised 26 January 2014; Accepted 10 February 2014; Published 2 April 2014

Academic Editor: Jiin Y. Jang

Copyright © 2014 S. M. Hashemi-Dehkordi et al. This is an open access article distributed under the Creative Commons
Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is
properly cited.

This paper presents friction-induced vibration (FIV) caused by combined mode-coupling and negative damping effects in a simple
FIV model. In doing so, a new four-degree-of-freedom linear model which consists of a slider and a block is proposed and then
simulated using MATLAB/Simulink. Stability or instability of the FIV model is defined by the convergence or divergence of
time domain responses of the slider and the block. Having found critical slope of friction-velocity characteristics that generate
instabilities in the model, a conventional closed loop proportional-integral-derivative (PID) controller is first introduced into the
main model in order to attenuate the vibration level and subsequently to suppress it. Later, the model is integrated with the active
force control (AFC) element to effectively reject the disturbance and reduce the vibrations. It is found that the integrated PID-AFC
scheme is effective in reducing vibration compared to the pure PID controller alone. Thus, the proposed control scheme can be one
of the potential solutions to suppress vibration in a friction-induced vibration system.

1. Introduction
Friction-induced vibration (FIV) exists in almost all mech-
anisms that are utilizing solid components that are rubbing
against each other, such as hard disks, disk brake system,
chain saw, and position control robots. The studies of friction-
induced vibration have been carried out over decades by
analytical [1–5], numerical [6–8], and experimental method
[9–12] but there is still no method available to completely
suppress the friction-induced vibration issue. Furthermore,
a complete understanding of the problem has still not been
achieved. Papinniemi et al. [13] suggested that these are
mainly due to the complexity of the mechanisms themselves
and competitive nature of automotive industry which limits
the amount of cooperative research that is published in the
open literature or public domain. It is well accepted that
there are three major mechanisms that contribute to the
generation of friction-induced vibration, namely, negative
damping [1, 2, 10], sprag/stick slip [3, 4, 14], and mode
coupling [5, 6]. Apart from these mechanisms, moving load
[15–17] is another mechanism that leads to friction-induced
vibration. A detailed explanation of these mechanisms can be
found in [18].
There are typically two different analysis methodologies
available to predict stability or instability of the FIV assembly,
namely, complex eigenvalue analysis and dynamic transient
analysis. Both methodologies have their pros and cons
discussed by Mahajan et al. [19] and Ouyang et al. [20].
It is common that complex eigenvalue analysis is used to
predict squeal conditions and squeal frequencies [6, 7, 20].
The positive real parts of the complex eigenvalues indicate a
potential unstable vibration mode of the linear model of a
disc brake and are thought to show the likelihood of squeal
occurrence or the noise intensity [6]. On the other hand,
instability in the friction-induced vibration system can be
associated with an initially divergent vibration response using
dynamic transient analysis.

Downloaded from ade.sagepub.com by guest on June 23, 2016

2 Advances in Mechanical Engineering
In analytical approach using a minimal model of a FIV
system assembly, North [5] was the first researcher who stud-
ied brake noise and vibration using a 2-degree-of-freedom
disk brake model with mode coupling effect and analysed
them using eigenvalue analysis. The friction coefficient was
assumed to be constant. The instabilities of the model were
indicated by the positive real parts of the eigenvalues. Yang
et al. [3] proposed a new theory of brake noise caused by FIV
called sprag slip, and the mathematical model consisted of a
pined bar whose end is on a horizontal plane; the bar makes
an angle less than 90∘ with the plane. As the plane moves
forward the bar sprag to it and then due to the flexibility
(stiffness) within the system it allowed the bar to free itself
from the spragging condition. Once the spragging has been
relieved, the original contact situation is reestablished. This
process continued and could lead to a sprag-slip limit cycle.
The unstable oscillation in the system could also occur even
with constant friction coefficient. Shin et al. [2] developed
two simple 2-degree-of-freedom models of a FIV system that
represented the mechanism of negative damping. In-plane
and out-of-plane vibration were described. They showed
that from complex eigenvalue analysis, the damping is the
most important parameter for in-plane vibration, whereas
the stiffness is the most important parameter for the out-of-
plane vibration. Hoffmann et al. [4] employed FIV model that
had 2-degree-of-freedom and it was based on the pad/block
with considering the mechanism of modal coupling. In their
work, through eigenvalue analysis, friction was considered as
a significant factor that caused instability on both in-plane
and out-of-plane vibration.
Currently, there are many techniques or methods that
have been put forward to reduce or eliminate FIV noise
and vibration. In a recent review, Chen et al. [21] pro-
posed guidelines to reduce and eliminate squeal events.
This included optimization of the damping, minimizing the
impulsive excitation, and reducing the modal coupling. These
three guidelines have been adopted by many researchers
and thought to be essential for squeal reduction methods.
Different techniques such as structural modifications and
control methods have also been employed by previous
researchers to reduce the effect of friction-induced vibra-
tion. Nakano and Maegawa [22] introduced safety-design
criteria to reduce friction-induced vibration by conducting
dimensionless analysis and numerical simulation for a single-
degree-of-freedom system with friction. The model consid-
ered a discontinuity between static and kinetic friction and
the dependence of the kinetic friction coefficient on the
relative velocity. Due to dimensionless description the num-
ber of parameters reduced from nine to five, where four of
the five dimensionless parameters controlled the occurrence
limit of friction-induced vibration. The occurrence-limit
equation was derived on the basis of a previous study on stick
slip in the Coulomb friction model, and the discriminant
inequalities were constructed with the four parameters in
which they are sufficient conditions for preventing friction-
induced vibration. Ouyang et al. [23] presented structural
modification approach in order to reduce vibration using
complex eigenvalue analysis, by determining the latent roots
of asymmetric systems represented by second order matrix
Downloaded from ade.sagepub.com by guest on June 23, 2016
differential equations and predicting the critical value of the
parameter generating the asymmetry in the stiffness matrix,
based on the reacceptance of the corresponding symmetric
systems. Chatterjee and Mahata [24] employed an active
absorber based on the time-delayed displacement difference
feedback in controlling friction-driven vibrations. The local
stability analysis shows that the static equilibrium can be
locally stabilized by suitably selecting the control gain and
the time delay. The regions of stability are delineated in the
plane of the control parameters. Numerical simulations of the
system show that proper selection of the control parameters
can also achieve the global stability of the system.
There is also a control method known as dither technique
[25–29] that has been introduced to prevent brake noise. The
dither technique makes use of high frequency disturbance
signal for the suppression of the automotive disc brake squeal
caused by FIV. In this scheme, the dither signal stabilizes
friction-induced self-oscillations in the disc brake using a
harmonic vibration, with a frequency higher than the squeal
frequency generated from a stack of piezoelectric elements
placed on the piston. The resulting control vibration was not
heard from the brake system if an ultrasonic control signal
was activated. This system is based on an open loop control
mode in which there is no requirement to detect the presence
of squeal.
It is found that a very few FIV models [1–30] include
the effects of mode coupling and negative damping together.
There are also not many FIV models for investigating the
vibration response of the disk/slider. Therefore, the present
work attempts to combine these two mechanisms together
into a new minimal four-degree-of-freedom friction-induced
vibration linear model. Stability or instability of the FIV
model is defined by the convergence or divergence of time
domain response of the slider and the block, respectively. In
order to prevent instability to the system, a control method
which is a closed loop control employing PID element with
and without active force control (AFC) is proposed and
integrated into the new FIV model. The main advantage of
the AFC technique is its ability to reject disturbance that is
applied to the system through appropriate manipulation of
the selected parameters. Additionally, the technique requires
much less computational burden and has been successfully
established to be readily implemented in real time.
2. A Simple 4-DOF Friction-Induced
Vibration Linear Model
In this paper a simple 4-DOF FIV linear model is proposed
that consists of a block and a slider as shown in Figure 1.
The model considers a combined effect of mode coupling and
negative damping. In the model, 𝑚1 and 𝑚2 represent mass of
the block and the slider, respectively. Each mass is capable of
moving in the in-plane (𝑥1 , 𝑥2 ) and out-of-plane directions
(𝑦1 , 𝑦2 ). Two diagonal springs are attached to the masses in
order to couple the equations of motion and hence they will
affect each other in the vibration response in both directions.
There is one spring 𝑘𝑐 that attaches between two masses to
represent contact stiffness. The friction force is generated due
Advances in Mechanical Engineering 3

The stiffness of the spring, 𝑘, the normal force, 𝑁, and the

𝜃 k2 N0 c2 friction force, 𝐹, are defined as follows:
k3
𝑁 = 𝑘𝑐 × (𝑦1 − 𝑦2 ) ,
k1 y1

m1 Vrel = (V𝑏 + 𝑥2̇ ) − 𝑥1̇ ,

x1
c1
𝑘11 = 𝑘3 cos2 𝜃 + 𝑘1 ,
kc
𝑘12 = 𝑘21 = −𝑘3 cos 𝜃 sin 𝜃,
b
k5 Ff y2 𝑘34 = 𝑘43 = 𝑘6 cos 𝜃 sin 𝜃, (3)
N
2
m2
𝑘22 = 𝑘3 sin 𝜃 + 𝑘2 ,
c3 x2

k6 c4
𝑘33 = 𝑘6 cos2 𝜃 + 𝑘5 ,
k4
𝜃 𝑘44 = 𝑘6 sin2 𝜃 + 𝑘4 ,

Figure 1: A simple 4-DOF FIV linear model. 𝐹𝑓 = 𝜇𝑘 × 𝑁,

where 𝜇𝑘 and 𝜇𝑠 are the dynamic and static friction coeffi-

𝜇k cients, respectively, Vrel is the relative velocity between the
𝜇s slider and the block, and 𝛼 is the negative slope of the linear
function of friction-velocity curve. The 𝜀 in (1) is a very small
positive number, and it is there to prevent the sign function
from becoming zero for a situation that the relative velocity is
o zero. Thus if the relative velocity is zero the friction coefficient
rel would be equal to the static friction coefficient (𝜇𝑠 ). It can be
seen from the model that by describing the dynamic friction
−𝜇s
coefficient as mentioned in (1) the effect of negative damping
is considered in the model, and with contact stiffness, 𝐾𝑐 , the
stiffness matrix in (2) becomes asymmetric and hence it leads
−𝜇k
to the effect of mode coupling.
Figure 2: Linear dynamic friction coefficient over relative speed.
3. Control Strategy
to the horizontal movement of the masses, through the slider After obtaining the mathematical model that represents the
and the contact spring that are connected to the masses. The effect of negative damping and mode coupling as shown in
slider that is connected to the mass 𝑚2 has an initial and Figure 1, it is necessary to control the vibration in all four
constant velocity of 𝑉𝑏 . 𝑁0 and 𝑁 are the preload force and the directions 𝑥1 , 𝑦1 , 𝑥2 , and 𝑦2 byusing actuators which can
normal reaction force of the surface, respectively. The preload apply forces that are parallel to the given axes. Thus the
force is necessary to excite the system. equations of motion having introduced the actuators can be
The definition of the dynamic friction coefficient is shown given as follows:
in Figure 2 and is given in ..
󵄨 󵄨 𝑚1 0 0 0 𝑥1 𝑐1 0 0 0 𝑥1̇
𝜇𝑘 = ((−𝛼 × 󵄨󵄨󵄨Vrel 󵄨󵄨󵄨) + 𝜇𝑠 ) × sgn (Vrel + 𝜀) . (1) [0 𝑚1 0 0]
..
𝑦̇
[ 𝑦
] ( ..1 ) + ( 0 𝑐2 0 0
) ( 1)
[0 0 𝑚2 0 ] 𝑥2 0 0 𝑐3 0 𝑥2̇
The equations of motion of the 4-DOF FIV model are ..
given as follows: [0 0 0 𝑚2 ] 𝑦2 0 0 0 𝑐4 𝑦2̇
..
𝑚1 0 0 0 𝑥1 𝑐1 0 0 0 𝑥1̇ 𝑘11 𝑘12 0 0 𝑥1
[0 𝑚1 0 0]
..
𝑦̇ [𝑘21 𝑘22 0 0 ] 𝑦1
[ 𝑦
] ( .. ) + ( 0
1 𝑐2 0 0
) ( 1) +[ ]
[ 0 0 𝑘33 𝑘34 ] (𝑥2 )
[0 0 𝑚2 0 ] 𝑥2 0 0 𝑐3 0 𝑥2̇
..
[0 0 0 𝑚2 ] 𝑦2 0 0 0 𝑐4 𝑦2̇ [ 0 0 𝑘34 𝑘44 ] 𝑦2

𝑘11 𝑘12 0 0 𝑥1 −𝐹𝑓 − (𝜇𝑘 × 𝑁0 ) −𝐹𝑓 − (𝜇𝑘 × 𝑁0 ) + 𝐴

[𝑘21 𝑘22 0 0 ] 𝑦1 [ − (𝑁 + 𝑁0 ) ] [ − (𝑁 + 𝑁0 ) + 𝐵 ]
+[ ] [ ]. =[ ],
[ 0 0 𝑘33 𝑘34 ] (𝑥2 ) = [ +𝐹𝑓 ] [ +𝐹𝑓 + 𝐶 ]
0 0 𝑘 𝑘 𝑦 +𝑁 [ +𝑁 + 𝐷 ]
[ 34 44 ] 2 [ ]
(2) (4)

Downloaded from ade.sagepub.com by guest on June 23, 2016

4 Advances in Mechanical Engineering

𝜃 k2 B N0 y1
c2
k3
k1 x1

m1 A
c1

kc y2
D b
k5 Ff x2
N
m2
c3 C
k6 c4
k4
𝜃

Figure 3: Actuator forces applied to the model.

Disturbance

+
+ + +
PID ẍ ẋ x
Reference Integrator Integrator Actual
Actuator Dynamic
input controller output
model (1/s) (1/s)
− +

Accelerometer
Force sensor

Estimated
mass
+ −

1/F Active force control

Displacement
transducer

Figure 4: Schematic diagram of AFC scheme.

where forces 𝐴, 𝐵, 𝐶, and 𝐷 are the actuator forces that are 𝑥2̈ = (−𝑐3 𝑥2̇ − (𝑘6 cos2 𝜃 + 𝑘5 ) 𝑥2
applied to the model as shown in Figure 3. Equation (4) is
−1
then derived into state space as follows: − (𝑘6 cos 𝜃 sin 𝜃) 𝑦2 + 𝐹𝑓 + 𝐶) × (𝑚2 ) ,

𝑥1̈ = (−𝑐1 𝑥1̇ − (𝑘3 cos2 𝜃 − 𝑘1 ) 𝑥1 𝑦2̈ = ( − 𝑐4 𝑦2̇ − (𝑘6 cos 𝜃 sin 𝜃) 𝑥2
−1
+ (𝑘3 cos 𝜃 sin 𝜃) 𝑦1 − 𝐹𝑓 − (𝜇𝑘 − 𝑁0 ) + 𝐴) × (𝑚1 ) ,
−1 − (𝑘6 sin2 𝜃 + 𝑘4 ) 𝑦2 + 𝑁 + 𝐷) × (𝑚2 ) .
(5)
𝑦1̈ = (−𝑐2 𝑦̇̈ 1 + (𝑘3 cos 𝜃 sin 𝜃) 𝑥1
The control strategy that is proposed here employs
−1
− (𝑘3 sin2 𝜃 + 𝑘2 ) 𝑦1 − (𝑁 + 𝑁0 ) + 𝐵) × (𝑚1 ) , an AFC-based scheme that is used in conjunction with

Downloaded from ade.sagepub.com by guest on June 23, 2016

Advances in Mechanical Engineering 5

0
+ PID ++ .5
−
Reference x1
input (X1 ) PID1 Actuator
X1 X1

X1 out x1
-K- +
− X1 in
Switch1 1/Q x1 Memory1 .4
X1 ddot out x2
EM x1 X2
0 + PID ++ .5
−
Reference PID2 Actuator X2 out
input (X2 ) X2
x2
X2 in
+
-K- −
X2 ddot out
Switch2 -K-
1/Q x2 Memory4
EM x2
0 + PID ++ .5 Y1 out y1
−
Reference PID3 Actuator Y1 in Y1
input (Y1 ) Y1
Y1 ddot out
-K- + y1
−
Switch3
1/Q y1 Memory3 -K- Y2 out
EM y1 Y2 in
y2
0 + .5 Y2 ddot out
− PID ++ Y2
Reference PID4 Actuator
input (Y2 ) Y2
Passive FIV
system
+ y2
-K- −
Switch4
1/Q y2 Memory2 .5
EM y2

Figure 5: Simulink diagram of the proposed AFC model.

the conventional PID controller. AFC as first proposed by
Hewit and Burdess [30] is robust and effective in controlling
a robot arm. Mailah and fellow researchers [31–36] demon-
strated the application of the technique to include many
other dynamical systems with the incorporation of artificial
intelligence (AI) methods. The PID controller was first tuned
with Ziegler-Nichol’s method and then manipulated for fine
performance. Afterwards, the AFC part is included into the
system to supply the compensation of the disturbance that is
inherent in the FIV system. Figure 4 shows the AFC scheme
applied to a dynamic translation system of the adopted
models. AFC scheme is shown to be effective in providing
the actuated force and body acceleration that are precisely
measured and at the same time the estimated mass suitably
approximated [37, 38]. The essential AFC equation can be
related to the computation of the estimated disturbance force,
𝐹𝑑 , as follows:
𝐹𝑑 = 𝐹 − (EM) (𝑎) , (6)
where 𝐹 is the measured actuating force, EM is the esti-
mated mass, and 𝑎 is the measured linear acceleration. This
parameter is then fed back through a suitable inverse transfer
function of the actuator to be summed up with the PID
control signal. The theoretical analysis with the stability of
the proposed AFC method has been described in [39]. A
number of methods to estimate the mass have been proposed
in the previous studies such as artificial intelligence (AI)
and crude approximation techniques [37, 39]. In this study,
the use of crude approximation method to approximate the
estimated mass is seen sufficient. The main challenge of

Downloaded from ade.sagepub.com by guest on June 23, 2016

6 Advances in Mechanical Engineering

×10−5 ×10−4
12 0

10

8
−1
6

y1 (m)
x1 (m)

2 −2
0

−2
−3
−4

−6
0 1 2 3 4 5 0 1 2 3 4 5
Time (s) Time (s)
(a) In-plane vibration of 𝑚1 (b) Out-of-plane vibration of 𝑚1
−4 −5
×10
2 4 ×10

2
1.5
0

1 −2

−4
x2 (m)

y2 (m)

0.5
−6

0 −8

−10
−0.5
−12

−1 −14
0 1 2 3 4 5 0 1 2 3 4 5
Time (s) Time (s)
(c) In-plane vibration of 𝑚2 (d) Out-of-plane vibration of 𝑚2

Figure 6: Vibration response of the passive FIV system with 𝛼 = 0.01 and 𝜃 = 60∘ .

the AFC method is to acquire suitable estimation of the mass the necessary external energy to suppress vibration in the
required to compute the disturbance 𝐹𝑑 in the feedback loop. model. The parameters used in this study were mostly taken
A conventional PID that is used with the AFC scheme can be from the previous researchers [40–43]. However, it needs to
typically represented by the following equation: be manipulated to suit the application in the simulation and
within MATLAB capabilities. The detailed parameters are
𝐾𝑖 given as follows.
𝐺𝑐 (𝑠) = 𝐾𝑝 + + 𝐾𝑑 𝑠, (7)
𝑠 Mathematical model parameters:

where 𝐾𝑝 , 𝐾𝑖 , and 𝐾𝑑 are the proportional, integral, and (i) body masses: 𝑚1 = 0.3 kg and 𝑚2 = 2 kg;
derivative gains, respectively. (ii) spring stiffness: 𝑘1 = 260000 N/m, 𝑘5 = 665000 N/m,
𝑘3 = 270000 N/m, 𝑘6 = 400000 N/m, 𝑘2 =
4. Simulation 115000 N/m, 𝑘4 = 950000 N/m, and 𝑘𝑐 = 85000 N/m;
(iii) damping coefficient: 𝑐1 = 1.5 Ns/m, 𝑐2 = 1.5 Ns/m,
In order to simulate the four-DOF FIV model with the con- 𝑐3 = 0.1 Ns/m, and 𝑐4 = 0.1 Ns/m;
trollers, MATLAB, Simulink, and Control System Toolbox
(CST) software are utilized. The actuators are assumed to be (iv) static friction coefficient: 𝜇𝑠 = 0.6;
of a linear type with a suitable constant gain. They provide (v) normal preload: 𝑁0 = 100 N;

Downloaded from ade.sagepub.com by guest on June 23, 2016

Advances in Mechanical Engineering 7

−4 −4
2.5 ×10 0 ×10

−0.5
2
−1
1.5
−1.5
x1 (m)

y1 (m)
1 −2

−2.5
0.5
−3
0
−3.5

−0.5 −4
0 1 2 3 4 5 0 1 2 3 4 5
Time (s) Time (s)
(a) In-plane vibration of 𝑚1 (b) Out-of-plane vibration of 𝑚1
−4 −4
5 ×10 1.5 ×10
4
1
3
2
0.5
1
x2 (m)

y2 (m)

0 0
−1
−0.5
−2
−3
−1
−4
−5 −1.5
0 1 2 3 4 5 0 1 2 3 4 5
Time (s) Time (s)
(c) In-plane vibration of 𝑚2 (d) Out-of-plane vibration of 𝑚2

Figure 7: Vibration response of the passive FIV system, 𝛼 = 0.03 and 𝜃 = 60∘ .

(vi) angle: (𝜃) = 60∘ , 45∘ ;
(vii) initial constant velocity of the disk: 𝑉𝑏 = 20 km/h;
(viii) actuators gain: 𝑄 = 0.5 and actuators power: 𝑃 = 7
Watt;
(ix) reference input = 0.00 m (i.e., no vibration).
In the simulation process, several types of operating
conditions are deliberately introduced to the mathematical
model to evaluate the effectiveness of the control system.
Four actuator forces for compensating the disturbance
that is actually inherent in the 4-DOF model are required.
The actuator forces are controlled by four individual PID
controllers involving four negative feedback loops.
For better vibration prevention, AFC is integrated into the
pure PID controller. The AFC Simulink diagram includes the
estimated mass, EM, and the parameter, 1/𝑄. The input to
the AFC is the mass acceleration and the output is summed
with the PID controller output and then multiplied with
the actuator gain which subsequently generates the actuator
force. To obtain the effective results, it is required to acquire a
suitable estimated mass combined with appropriate tuning of
the PID controller gains. Figure 5 shows an AFC scheme used
in the study to suppress vibration and it can be seen that each
PID controller is added to the AFC loop. It can be noticed that
four AFC loops are added to the four PID controllers.
In order to tune the PID controllers, Ziegler-Nichol’s
method is used with manipulation to suit the system perfor-
mance and the results are tabulated as shown in Table 1. Also,
it shows the values of the estimated mass in which they were
mainly obtained by trial and error or crude approximation
method. It should be mentioned that the percentage of the
AFC used is 100% (switch is set to ON).

Downloaded from ade.sagepub.com by guest on June 23, 2016

8 Advances in Mechanical Engineering

×10−5 −4
10 0 ×10

−0.5
8
−1
6
−1.5
4 −2

y1 (m)
x1 (m)

2 −2.5

−3
0
−3.5
−2
−4

−4 −4.5
0 1 2 3 4 5 0 1 2 3 4 5
Time (s) Time (s)
(a) In-plane vibration of 𝑚1 (b) Out-of-plane vibration of 𝑚1
−4 −4
2 ×10 1 ×10

1.5 0.5

1
0
0.5
x2 (m)

y2 (m)

−0.5
0
−1
−0.5

−1 −1.5

−1.5 −2
0 1 2 3 4 5 0 1 2 3 4 5
Time (s) Time (s)
(c) In-plane vibration of 𝑚2 (d) Out-of-plane vibration of 𝑚2

Figure 8: Vibration response of the passive FIV system with 𝛼 = 0.03 and 𝜃 = 45∘ .

5. Results and Discussion Table 1: The values for the estimated mass and the PID parameters.

In this section the results of the four-degree-of-freedom PID parameters 𝐾𝑝 𝐾𝑖 𝐾𝑑

model of friction-induced vibration, which represents both PID 𝑥1 4 1.5 13.5
effects of negative damping and modal coupling. At first in PID 𝑦1 5 2 11.2
Section 5.1 in order to approve the correctness and appro-
PID 𝑥2 5 1.8 15
priate operation of the model some simulation tests were
executed from the model when it did not have any controllers PID 𝑦2 7 3 18
applied to it (i.e., the passive model). Then in Section 5.2 Estimated mass (kg)
simulation tests were carried out from the 4-DOF model that EM 𝑥1 0.4
was equipped with PID and AFC controllers. EM 𝑦1 0.45
EM 𝑥2 0.6
5.1. Passive 4-DOF Model. By giving different values of the EM 𝑦2 0.5
slope and considering the angle (𝜃) to be 60∘ [41], a number
of simulations were performed using the passive FIV system,
that is, without any controller attached to it. These simulation
tests are executed by giving different values of the slope (𝛼) are shown in Figure 6. It can be seen that the response of the
in (1). First, a value of 𝛼 = 0.01 was given and the results block (𝑚1 ) and the slider (𝑚2 ) in both in-plane (𝑥1 and 𝑥2 )

Downloaded from ade.sagepub.com by guest on June 23, 2016

Advances in Mechanical Engineering 9

×10−5 ×10−4
16 0.5
14
0
12
10 −0.5

8 PID PID + AFC

−1
x1 (m)

y1 (m)
6
PID + AFC PID
4 −1.5

2 −2
0
−2.5
−2
−4 −3
0 1 2 3 4 5 0 1 2 3 4 5
Time (s) Time (s)
(a) In-plane vibration of 𝑚1 (b) Out-of-plane vibration of 𝑚1

×10−4 −5
2 ×10

0
2
−2
1.5
PID −4
PID + AFC
x2 (m)

y2 (m)

1 −6
PID + AFC
−8 PID
0.5
−10
0
−12

−0.5 −14
0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 0 1 2 3 4 5
Time (s) Time (s)
(c) In-plane vibration of 𝑚2 (d) Out-of-plane vibration of 𝑚2

Figure 9: Vibration response of the FIV model with 𝛼 = 0.03 and 𝜃 = 60∘ .

and out-of-plane (𝑦1 and 𝑦2 ) directions under the given value
of the slope is converging and it is indicated that the system is
under stable condition. Due to the reason that the value of the
negative damping slope is considerably low, the changes of the
friction coefficient with the increase of the relative velocity are
superficial and thus the system is stabilizing with time. The
offset that can be seen in the out-of-plane results is mainly
due to the natural behavior of the system. The preload 𝑁𝑜 will
initially press the block and slider down until they reach the
equilibrium condition.
Later the slope (𝛼) was increased to 0.03 and time domain
results are shown in Figure 7. It can be seen that the block
(𝑚1 ) and the slider (𝑚2 ) are clearly under unstable condition
in the out-of-plane direction and in-plane direction. Because
the slope of negative damping is higher than the previous case
it causes the friction coefficient to decrease with a noticeable
amount with the increase of relative velocity, and thus the
response of the system goes toward instability.
One more simulation was performed by considering the
angle 𝜃 to be 45∘ in order to examine its influence on the
system stability. The slope was set to 0.03 and the results of
this simulation are shown in Figure 8. It can be seen that when
the angle 𝜃 is set to 45∘ both 𝑚1 and 𝑚2 are less diverging,
compared to the case where the angle was set to 60∘ . It is also
found that when the angle 𝜃 is decreased from 60∘ to 45∘ ,
the value of the slope that is required to make the FIV model
unstable is higher.
5.2. Reducing the Vibrations of the 4-DOF Model with AFC.
After studying behavior of the four-DOF passive FIV model,
it is necessary to reduce and eliminate the vibration in it. At
first the simulation is performed at the angle 𝜃 of 60∘ with

Downloaded from ade.sagepub.com by guest on June 23, 2016

10 Advances in Mechanical Engineering

×10−4 ×10−4
5 1

PID
4 0

3 −1
PID + AFC
x1 (m)

y1 (m)
2 −2

PID + AFC
1 −3

0 −4 PID

−1 −5
0 1 2 3 4 5 0 1 2 3 4 5
Time (s) Time (s)
(a) In-plane vibration of 𝑚1 (b) Out-of-plane vibration of 𝑚1
−3 −4
1 ×10 2 ×10
0.8 1.5
PID + AFC PID + AFC
0.6
1
0.4
0.2 0.5
x2 (m)

y2 (m)

0 0
−0.2 −0.5
−0.4
0.4
−1
−0.6 PID PID
−1.5
−0.8
−1 −2
0 1 2 3 4 5 0 1 2 3 4 5
Time (s) Time (s)
(c) In-plane vibration of 𝑚2 (d) Out-of-plane vibration of 𝑚2

Figure 10: Vibration response of the FIV model with 𝛼 = 0.08 and 𝜃 = 60∘ .

the slope of 0.03 and results of this simulation are shown
in Figure 9. It can be seen that when AFC is applied to the
model, in-plane and out-of-plane vibration of the block (𝑚1 )
and the slider (𝑚2 ) are better reduced compared to the case
that the model has only PID controllers.
By increasing the slope to 0.08 the results show that the
PID controllers are not effective in reducing the vibrations
and the FIV model is diverging, but by engaging the AFC loop
into the PID controllers the vibrations are very effectively
reduced to nearly zero as shown in Figure 10. This shows
that a PID plus AFC controller is very effective in preventing
vibration for a quite large range of the slope.
In order to evaluate the robustness of the system the
simulation was executed with having the mass 𝑚1 reduced
by 10%. The results of this simulation are shown in Figure 11,
and it can be seen that AFC is still capable of reducing the
vibrations to a noticeable amount.
Another simulation was executed with considering a non-
linear relation between the friction coefficient and relative
velocity [44]; in this case the friction coefficient is considered
to be in a sinusoidal format. The results of this simulation
are represented in Figure 12 and they show that after AFC
is engaged into the system the noise or vibration reduces to
a suitable manner. The out-of-plane vibrations showed the
same behavior.
6. Conclusion
A four-degree-of-freedom linear model of a friction-induced
vibration system has been developed, in which both mech-
anisms of negative damping and mode coupling which are
two of the most important mechanisms that cause friction-
induced vibration are represented. An active force control-
based method has been developed and implemented to

Downloaded from ade.sagepub.com by guest on June 23, 2016

Advances in Mechanical Engineering 11

×10−3 3 ×10
−3
2
PID + AFC
1.5
2
1 PID + AFC
1
0.5
x1 (m)

x2 (m)
0 0

−0.5
−1
−1 PID
−2
−1.5 PID
−2 −3
0 1 2 3 4 5 0 1 2 3 4 5
Time (s) Time (s)
(a) In-plane vibration of 𝑚1 (b) In- plane vibration of 𝑚2

Figure 11: Vibration response of the FIV model with 𝛼 = 0.08 and 𝜃 = 60∘ and the mass 𝑚1 reduced by 10%.

−4 −5
0.5 ×10 2 ×10

0
0
−0.5
PID + AFC −2
−1 PID + AFC
−1.5 −4
x2 (m)
x1 (m)

−2 −6
−2.5 PID
−8 PID
−3
−10
−3.5

−4 −12
0 1 2 3 4 5 0 1 2 3 4 5
Time (s) Time (s)
(a) In-plane vibration of 𝑚1 (b) In-plane vibration of 𝑚2

Figure 12: Vibration response of the FIV model with 𝛼 = 0.08 and 𝜃 = 60∘ , with a nonlinear sinusoidal friction coefficient.

significantly reduce the vibration caused due to the effects of
negative damping and mode coupling. The AFC component
is readily incorporated into a PID controller and the resulting
four-DOF scheme has been shown to be very effective in
suppressing the vibration in the 4-DOF FIV model for a
number of different operating conditions. It was observed
that, in some cases, a pure PID controller alone is able to
reduce the vibration but more often than not it is much less
effective compared to the PID plus AFC scheme.
With the increase in the slope of the linear friction
coefficient in negative damping, a pure PID controller was
unable to stabilize the system and the vibration was shown
to gradually increase in amplitude with time. However, when
the AFC mode is activated, the vibration was very much
reduced, thereby rendering the system very effective. It is
worth to see whether the AFC mode is capable of suppressing
vibration if nonlinear aspects of the surface contact and of
mu-velocity slope are considered. This is the authors’ recent
work. Due to the equipment and the required hardware that
are used for the AFC system, the cost of this method at the
moment is expected to be expensive and thus it can be used
as a luxury option or in sensitive cases.
Conflict of Interests
The authors declare that there is no conflict of interests
regarding the publication of this paper.

Downloaded from ade.sagepub.com by guest on June 23, 2016

12
References
[1] Y. Yuan, “A study of the effects of negative friction-speed
slope on brake squeal,” in Proceedings of the 1995 ASME
Design Engineering Technical Conference, vol. 84-1, pp. 1153–
1162, September 1995.
[2] K. Shin, M. J. Brennan, Y.-G. Joe, and J.-E. Oh, “Simple models
to investigate the effect of velocity dependent friction on the
disk brake squeal noise,” International Journal of Automotive
Technology, vol. 5, no. 1, pp. 61–67, 2004.
[3] F. H. Yang, W. Zhang, and J. Wang, “Sliding bifurcations and
chaos induced by dry friction in a braking system,” Chaos,
Solitons and Fractals, vol. 40, no. 3, pp. 1060–1075, 2009.
[4] N. Hoffmann, M. Fischer, R. Allgaier, and L. Gaul, “A minimal
model for studying properties of the mode-coupling type
instability in friction induced oscillations,” Mechanics Research
Communications, vol. 29, no. 4, pp. 197–205, 2002.
[5] M. R. North, “Disc brake squeal—a theoretical model,” Tech.
Rep. 1972/5, MIRA, 1972.
[6] G. D. Liles, “Analysis of disc brake squeal using finite element
methods,” S.A.E. Paper 891150, 1989.
[7] D. Guan and D. Jiang, “A study on disc brake squeal using finite
element methods,” S.A.E. Technical Paper 980597, 1998.
[8] A. R. Abu-Bakar, Modeling and simulation of disc brake contact
analysis and squeal [Ph.D. thesis], Department of Engineering,
University of Liverpool, Liverpool, UK, 2005.
[9] R. A. Fosberry and Z. Holubecki, “Disc brake squeal its
mechanism and suppression,” Tech. Rep. 19612, MIRA, 1961.
[10] J. D. Fieldhouse and P. Newcomb, “The application of holo-
graphic interferometry to the study of disc brake noise,” S.A.E.
Technical Paper 930805, 1993.
[11] J. D. Fieldhouse and T. P. Newcomb, “An investigation into
disc brake noise using holographic interferometry,” IMechE
C42711213, 1991.
[12] A. Akay, J. Wickert, and Z. Xu, “Investigating mode lock-in
and friction interface,” Final Research Report, Carnegie Mellon
University, 2000.
[13] A. Papinniemi, J. C. S. Lai, J. Zhao, and L. Loader, “Brake squeal:
a literature review,” Applied Acoustics, vol. 63, no. 4, pp. 391–400,
2002.
[14] S. W. E. Earles and C. K. Lee, “Instability arising from the
friction interaction of a Pin-Disc system resulting in noise
generation,” Journal of Engineering for Industry, vol. 98, no. 1,
pp. 81–86, 1976.
[15] J. E. Mottershead, H. Ouyang, M. P. Cartmell, and M. I. Friswell,
“Parametric resonances in an annular disc, with a rotating
system of distributed mass and elasticity; and the effects of
friction and damping,” Proceedings of the Royal Society A, vol.
453, no. 1956, pp. 1–19, 1997.
[16] H. Ouyang, J. E. Mottershead, M. P. Cartmell, and M. I. Friswell,
“Friction-induced parametric resonances in discs: effect of a
negative friction-velocity relationship,” Journal of Sound and
Vibration, vol. 209, no. 2, pp. 251–263, 1998.
[17] H. Ouyang, J. E. Mottershead, D. J. Brookfield et al., “Dynamic
instabilities in a simple model of a car disc brake,” S.A.E. Paper
1999-01-3409, 1999.
[18] N. M. Kinkaid, O. M. O’Reilly, and P. Papadopoulos, “Automo-
tive disc brake squeal,” Journal of Sound and Vibration, vol. 267,
no. 1, pp. 105–166, 2003.
[19] S. K. Mahajan, Y. K. Hu, and K. Zhang, “Vehicle disc brake
squeal simulations and experience,” S.A.E. Technical Paper
1999-01-1738, 1999.
Downloaded from ade.sagepub.com by guest on June 23, 2016
Advances in Mechanical Engineering
[20] H. Ouyang, W. Nack, Y. Yuan, and F. Chen, “Numerical analysis
of automotive disc brake squeal a review,” International Journal
of Vehicle Noise and Vibration, vol. 1, no. 34, pp. 207–230, 2005.
[21] F. Chen, C. A. Tan, and R. L. Quaglia, “Disk brake squeal
mechanism, analysis, evaluation and reduction prevention,”
S.A.E. Paper, 2006.
[22] K. Nakano and S. Maegawa, “Safety-design criteria of sliding
systems for preventing friction-induced vibration,” Journal of
Sound and Vibration, vol. 324, no. 3-5, pp. 539–555, 2009.
[23] H. Ouyang, L. Baeza, and S. Hu, “A receptance-based method
for predicting latent roots and critical points in friction-induced
vibration problems of asymmetric systems,” Journal of Sound
and Vibration, vol. 321, no. 3-5, pp. 1058–1068, 2009.
[24] S. Chatterjee and P. Mahata, “Time-delayed absorber for
controlling friction-driven vibration,” Journal of Sound and
Vibration, vol. 322, no. 1-2, pp. 39–59, 2009.
[25] A. J. Graf, Active control of automotive disc brake rotor squeal
using dither [M.S. thesis], Georgia Institute of Technology, 2000.
[26] K. A. Cunefare and A. J. Graf, “Disc brake rotor squeal
suppression using dither control,” SAE Paper 2001-01-1605,
2001.
[27] K. A. Cunefare and A. J. Graf, “Experimental active control
of automotive disc brake rotor squeal using dither,” Journal of
Sound and Vibration, vol. 250, no. 4, pp. 579–590, 2002.
[28] F. Chen, A. Wang, and C. A. Tan, “Suppression of self-excited
vibration by dither technique with potential application to
reduce brake squeal,” S.A.E. Paper 2004-01-2790, 2004.
[29] N. Hoffmann, N. Wagner, and L. Gaul, “Quenching mode-
coupling friction-induced instability using high-frequency
dither,” Journal of Sound and Vibration, vol. 279, no. 1-2, pp. 471–
480, 2005.
[30] J. R. Hewit and J. S. Burdess, “Fast dynamic decoupled con-
trol for robotics, using active force control,” Mechanism and
Machine Theory, vol. 16, no. 5, pp. 535–542, 1981.
[31] E. Pitowarno and M. Mailah, “Control of mobile manipulator
using resolved acceleration iterative learning proportional-
integral active force control,” International Review of Mechanical
Engineering, vol. 1, no. 5, pp. 549–558, 2007.
[32] G. Priyandoko, M. Mailah, and H. Jamaluddin, “Vehicle active
suspension system using skyhook adaptive neuro active force
control,” Mechanical Systems and Signal Processing, vol. 23, no.
3, pp. 855–868, 2009.
[33] S. M. Hashemi-Dehkordi, M. Mailah, and A. R. Abu Bakar,
“Intelligent active force control with piezoelectric actuators to
reduce friction induced vibration due to negative damping,”
International Review of Electrical Engineering, vol. 4, no. 6, pp.
1294–1305, 2009.
[34] S. M. Hashemi-Dehkordi, M. Mailah, and A. R. Abu-Bakar,
“Suppressing friction-induced vibration due to negative damp-
ing and mode coupling effects using active force control,”
Australian Journal of Basic and Applied Sciences, vol. 4, no. 8,
pp. 3917–3933, 2010.
[35] S. M. Hashemi-Dehkordi, M. Mailah, and B. A. R. Abu, “Imple-
mentation of active force control to disc brake for noise-free
performance,” International Review of Mechanical Engineering,
vol. 3, no. 4, pp. 481–488, 2009.
[36] S. M. Hashemi-Dehkordi, A. R. Abu-Bakar, and M. Mailah,
“Reducing friction-induced vibration using intelligent active
force control (AFC) with piezoelectric actuators,” Sadhana, vol.
37, part 6, pp. 637–655, 2012.
Advances in Mechanical Engineering 13

[37] M. Mailah, Intelligent active force control of a rigid robot arm

using neural network and iterative learning algorithms [Ph.D.
thesis], University of Dundee, Dundee, UK, 1998.
[38] M. Mailah and N. I. A. Rahim, “Intelligent active force control
of a robot arm using fuzzy logic,” in Proceedings of the IEEE
International Conference on Intelligent Systems and Technologies,
vol. 2, pp. I-291–I-296, Kuala Lumpur, Malaysia, September
2000.
[39] J. S. Burdess and J. R. Hewit, “An active method for the control of
mechanical systems in the presence of unmeasurable forcing,”
Mechanism and Machine Theory, vol. 21, no. 5, pp. 393–400,
1986.
[40] K. Shin, M. J. Brennan, J.-E. Oh, and C. J. Harris, “Analysis of
disc brake noise using a two-degree-of-freedom model,” Journal
of Sound and Vibration, vol. 254, no. 5, pp. 837–848, 2002.
[41] N. Hoffmann, S. Bieser, and L. Gaul, “Harmonic balance and
averaging techniques for stick-slip limit-cycle determination
in mode-coupling friction self-excited systems,” Technische
Mechanik, vol. 24, pp. 185–197, 2004.
[42] U. von Wagner, D. Hochlenert, and P. Hagedorn, “Minimal
models for disk brake squeal,” Journal of Sound and Vibration,
vol. 302, no. 3, pp. 527–539, 2007.
[43] H. Ouyang, “Prediction and assignment of latent roots of
damped asymmetric systems by structural modifications,”
Mechanical Systems and Signal Processing, vol. 23, no. 6, pp.
1920–1930, 2009.
[44] A. J. McMillan, “A non-linear friction model for self-excited
vibrations,” Journal of Sound and Vibration, vol. 205, no. 3, pp.
323–335, 1997.

Downloaded from ade.sagepub.com by guest on June 23, 2016