Tribology International 92 (2015) 418424

Contents lists available at ScienceDirect

Tribology International
journal homepage: www.elsevier.com/locate/triboint

A statistical model for evaluating the tribological properties of paperbased friction materials
Wenbin Li a, Jianfeng Huang a,n, Jie Fei a, Liyun Cao a, Chunyan Yao b, Wenjing Wang a
a
b

School of Materials Science & Engineering, Shaanxi University of Science and Technology, Xi'an, Shaanxi 710021, PR China
Culture and Communication School, Shaanxi University of Science and Technology, Xi'an, Shaanxi 710021, PR China

art ic l e i nf o

a b s t r a c t

Article history:
Received 16 October 2014
Received in revised form
19 June 2015
Accepted 23 July 2015
Available online 31 July 2015

In order to synthetically evaluate the tribological properties of friction materials, fuzzy comprehensive
evaluation model was developed. The analytic hierarchy process was introduced into the model to
establish weight function. The paper-based friction materials with different CNTs content were chosen as
the six projects. It is found that the comprehensive evaluation indexes are increased by 126307% for the
samples with 415 wt% of CNTs compared with the sample without CNTs. The judgment matrix has high
consistency by consistency checking. And the evaluation results are also in accordance with the results
obtained through the friction torque, surface structure and temperature.
& 2015 Elsevier Ltd. All rights reserved.

Keywords:
Fuzzy comprehensive evaluation model
Weight function
Analytic hierarchy process
Tribological properties

1. Introduction
As a kind of important ber-reinforced composites, paper-based
friction materials have been widely used for the wet clutches of
automatic transmissions because of their outstanding mechanical and
tribological properties. A large number of researches have been
conducted to better explore paper-based friction materials. Yi et al.
investigated the mechanical properties of phenolic resin-based friction composites and found that the bending strength and hardness
increase with the rise of calcined petroleum coke content [1]. Fujii
et al. studied the fatigue strength of aramid ber reinforced paperbased friction materials under shear-compressive loading and found
that the fatigue strength is remarkably high under the compressivedominant stress condition [2]. Wang et al. explored the wear
performance of carbon ber reinforced nylon 1010 composites from
the perspective of ller and got the nding that MoS2 ller increases
their wear [3]. Hwang et al. investigated the friction performance of
paper-based friction materials containing carbon nanotubes (CNT)
from the perspective of friction modier, nding that CNT modier
increases the fade resistance and friction stability [4]. Patnaik et al.
studied the fade resistance and recovery properties of ber reinforced
hybrid phenolic composites from the perspective of reinforcement
and found that the fade resistance and recovery increase with the
increase of aramid ber and decrease of ceramic ber [5]. Fei et al.
explored the tribological properties of carbon ber reinforced paperbased friction materials from the perspective of binder and discovered

Corresponding author. Tel./fax: 86 029 86168802.

E-mail address: huangjf@sust.edu.cn (J. Huang).

http://dx.doi.org/10.1016/j.triboint.2015.07.025
0301-679X/& 2015 Elsevier Ltd. All rights reserved.

that the samples with lower resin content (ranging from 35 wt% to
40 wt%) exhibit high friction coefcient, excellent friction stability,
good heat-resistance and reasonable strength [6]. Regarded as a kind
of important properties of paper-based friction materials, tribological
properties have been widely studied, such as dynamic friction
coefcient, the ratio of dynamic and static friction coefcient, static
friction coefcient, variation coefcient, wear rate and friction torque.
The dynamic friction coefcient can well reect the torque
transmission ability and braking efciency of paper-based friction
materials [7]. The static friction coefcient mainly exhibits the torque
transmissions ability in the later stage of shifting/joint and the high
static friction coefcient easily leads to serious shudder phenomenon.
The ratio of dynamic and static friction coefcient can demonstrate
the potential for friction-induced vibrations and the large ratio
indicates excellent anti-shudder performance [8]. The variation coefcient can reect the friction stability and the lower variation
coefcient indicates better friction stability [9]. The wear rate is the
key index characterizing the life of paper-based friction materials and
the small wear rate means long service life. Therefore, the above ve
indexes are very important to evaluate the tribological properties of
paper-based friction materials. However, these indexes are usually
inconsistent and show the complex nonlinear interaction, which
makes it difcult to determine whether a friction material that has
excellent tribological properties.
Kim et al. studied the friction stability and wear resistance of
friction materials containing two different phenolic resins [10]. It
was found that the friction materials with the modied novolac
resin show improved friction stability while reducing wear resistance. Fei et al. investigated the effects of carbon ber length (100,
400, 600 and 800 m) on the tribological properties of paper-

W. Li et al. / Tribology International 92 (2015) 418424

based friction materials. It was found that bigger porosity and

more covering of resin make the friction material with 600 m
carbon bers exhibit greater dynamic friction coefcient, but the
wear rate is also greater [11]. Kim et al. studied the friction and
wear performances of chopped glass bers reinforced friction
materials [12]. It was found that the chopped glass bers decrease
the wear rate at 100 1C, but the friction coefcient and friction
stability are also decreased. Zhang et al. explored the inuences of
compound mineral ber content (0, 5, 10, 15 and 20 wt%) on the
tribological properties of paper-based friction materials [9]. As was
found, the friction material with 15 wt% of compound mineral
ber has the smallest wear rate and variable coefcient, but the
dynamic friction coefcient is not the largest. Thus, in order to
settle the above inconsistency problems of multiple indexes,
comprehensive evaluation model is very necessary in that it can
unite multiple evaluation indexes into one index.
A number of evaluation methods have been employed in the
evaluation of friction materials. Mustafa et al. accomplished the
selection of friction materials with cambridge engineering selector
edupack software based on eco-aware lightweight, cost effective and
non-toxic [13]. However, the tribological properties were not considered. Dadkar et al. conducted the tribo-evaluation of yash-lled
and aramid bre reinforced hybrid polymer matrix composites by
linear polynomial approach and modied Rhee equation [14].
However, the evaluation method is only suitable for the situation
where independent variable and dependent variable can be measured. The selection and verication of kenaf bers as an alternative
friction material were conducted using weighted decision matrix
method by Mustafa et al. [15]. However, the determination of weight
factor is subjective relatively and the verication of evaluation result
is not abundant relatively. In order to comprehensively evaluate the
anti-shudder performance of ATFs, oil absorption ability of the posttest friction material surface was proposed to be an important
parameter [16]. Mortazavi et al. developed a global statistical
approach for the evaluation of mechanical properties of silica/epoxy
nanocomposite [17]. However, the global statistical approach cannot
unite multiple evaluation indexes into one index. Therefore, few
good approaches in previous researches were found to be suitable
for the evaluation of tribological properties.
However, the advantages of being concise, practical, systematic
and quantitative have enabled analytic hierarchy process (AHP) to be
widely used in consensus building [18] and ranking insurance rms
[19] and so on, but not in the evaluation of tribological properties.
Moreover, AHP can unite multiple evaluation indexes into one index.
Thus, AHP is suitable for evaluating the tribological properties of
paper-based friction materials based on the above analysis.
In the paper, the fuzzy comprehensive evaluation model was
developed based on AHP and weight function. Subsequently, it was
applied to the evaluation of the tribological properties of the paperbased friction materials with different CNTs content and the validity was
checked by the friction torque, surface structure and temperature. The
model can be well applied to the evaluation of the tribological properties
of paper-based friction materials, because it can well unite multiple
inconsistent evaluation indexes into one index. Moreover, the model can
help the designer more effectively, accurately and comprehensively
investigate paper-based friction materials and then select the friction
materials possessing excellent tribological properties. The model can also
provide evidence for material ratio and process optimization.

2. Methods and materials

2.1. Materials
The mixture of PAN short carbon bers (75150 m length,
supplied by Anshan Sinocarb Carbon Fibers Co., Ltd., Anshan,

419

China) and bamboo bers (901SR freeness, supplied by paper

making engineering department of Shaanxi University of Science
& Technology) was used as reinforcement. The friction modier
was CNTs (dispersed in water with styrenemaleic anhydride
copolymer (SMA), provided by the reaction engineering laboratory
of Tsinghua University). The PF-6291A cashew-modied phenolic
resin (provided by Shandong Shengquan chemical Co., Ltd., Jinan,
China) was used as binder. The ingredients used in this study are
listed in Table C1.
2.2. Fabrication of composites
The stable suspensions were obtained by mixing the carbon
bers, bamboo bers and CNTs in water and stirring for about
30 min. Then the preform sheets were shaped by pouring the
suspension into paper making machine. Afterwards, the preform
sheets were rolled and dried at 110 1C for 10 min. After that, the
dried preform sheets were dipped into the cashew-modied
phenolic resin solution (dissolved in the ethanol with the mass
concentration of 30%) for 30 min. After inltration, the preform
sheets were dried at room temperature, followed by compression
molding at 160 1C for about 5 min under the pressure of 5.0 MPa.
Thus, the paper-based friction materials were obtained, whose
thicknesses were 0.60.7 mm. The as-prepared composites containing 0, 2, 4, 8, 12 and 15 wt% of CNTs were designated as C1, C2,
C3, C4 and C5, respectively (Table C1). Finally, the test specimens
were cut from the as-prepared composites with the help of the
cutting mould according to the required standard (GB/T 138262008) for tribological properties test, whose schematic diagram is
shown in Fig. B1. And subsequently, the facings were adhesively
bonded to the supporting metal member under heat and pressure
in order to obtain the desired paper-based friction plate samples.
2.3. Tribological properties test
The tribological properties of the specimens were obtained by the
QM1000-II wet friction performance tester with plate-on-plate
conguration. Fig. B2 shows a schematic diagram of the equipment.
The as-prepared composites have an outside radius of 51.5 mm and
inner radius of 36.5 mm. The temperature and ow rate of lubrication oil (No. N32 engine oil) were kept at 40 1C and 90 ml/min in all
tribological properties tests. The temperatures of mating plate were
measured by thermometer (Fig. B2(12)) during each engagement and
then the data were transferred to the controller (Fig. B2(14)) where
there was a temperature indicator. Afterwards, the highest temperature was recorded by manual method. The measurements of
dynamic friction coefcient and static friction coefcient were
repeated for six times. And then an average value was calculated as
the result.
The change of sample thickness was measured after 500
engagement cycles under interface pressure of 1.0 MPa, initial
rotating speed of 2000 rpm and total inertia of 0.1294 Kg m2,
respectively. The wear rate was obtained from Eq. (A.1), which
was dened as wear volume per work and used for further
analysis, where V is wear rate, mm3/J; A is the apparent contact
area, mm2; h is the change of sample thickness, mm; n is the
number of engagement cycles; I 0 is the total inertia, Kg m2; is
the relative angular velocity, rad/s.
In this work, the variation coefcient was used to reect the
friction stability of the dynamic friction coefcient during 500
engagement cycles under interface pressure of 1.0 MPa, initial
rotating speed of 2000 rpm and total inertia of 0.1294 Kg m2,
respectively. The variation coefcients of the samples were calculated by Eq. (A.2), where C:V is the variable coefcient; is the
standard deviation; m is the average value of the dynamic friction
coefcient of 500 engagement cycles.

420

W. Li et al. / Tribology International 92 (2015) 418424

2.4. Microstructure test

Scanning electron microscope (SEM, S-4800) was operated at
3 keV to observe the morphologies of the paper-based friction
materials.

3. Establishment of fuzzy comprehensive evaluation model

Given the requirement of GB/T 13826-2008 and expert judgment, the paper-based friction materials cannot be used in the
engagement, when the dynamic friction coefcient is less than
0.08, the ratio of dynamic and static friction coefcient is greater
than 1 or the wear rate is greater than 6  10  2 mm3/J. Based on
the above analysis, weight function is constructed with 19 scale
method, expert judgment, data normalization and matrix solution
taken into account, which is the core idea of the fuzzy comprehensive evaluation model.
Firstly, through comprehensive analysis of the tribological
properties, ve criteria are selected and hierarchical structure
model is established. Secondly, with 19 scale method and expert
judgment used, the judgment matrix is constructed. The maximum eigenvector is used as the weight vector of the criteria. The
maximum eigenvalue is used for consistency checking. Finally, the
weight function is established based on the weights of the criteria.
And then the normalized data are put into the weight function to
obtain the value of comprehensive evaluation index, as shown in
Eqs. (A.3)(A.5), where M i is the average of index i for all specimens; xij is the test value of index i and specimen Pj; xij is the
normalized value of index i and specimen Pj; Y j is the comprehensive evaluation index value of specimen Pj.
In this paper, dynamic friction coefcient x1 , the ratio of
dynamic and static friction coefcient x2 , static friction coefcient x3 , variation coefcient x4 and wear rate x5 are chosen as
the criteria, which are viewed very important in measuring the
tribological properties of paper-based friction materials. Moreover,
it is necessary for x1 and x2 to be higher and x3 , x4 and x5 to be
lower in order to obtain the better tribological properties according to GB/T 13826-2008 and the studies of Fei et al. [20] and Zhang
et al. [9].

4. Results and discussion

4.1.3. Calculation of comprehensive evaluation index

The values of comprehensive evaluation indexes are calculated
by putting the weights of the criteria and the normalized date into
Eq. (A.5). Table C4 shows the values of comprehensive evaluation
indexes for the six projects. As shown in Table C4, P3 (4 wt% of
CNTs) presents excellent tribological properties for the CNTs
modied paper-based friction materials. In contrast, the P2 (2 wt
% of CNTs) presents poor tribological properties. It is also found
that the tribological properties of the paper-based friction materials with more than 4 wt% of CNTs added are superior to that of the
paper-based friction materials not containing CNTs. Chang et al.
found that the CNT increases the tribological properties of short
carbon/CNT/PA6 hybrid composites [21] and Hwang et al. found
that the CNT-added friction materials shows increased fade
resistance and friction stability [4]. Thus, the results are in
accordance with the reports of Chang et al. and Hwang et al.
Application of the model in different operating conditions
The tribological properties are affected by operating conditions,
especially interface pressure and rotating speed. Thus, the
dynamic friction coefcient changes with the variation of operating conditions [22]. As shown in Figs. B4 and B5, the dynamic
friction coefcient decreases with the increase of interface pressure and rotating speed. The mechanical interlocking of asperities
and the oil lm shear stress decrease with the increase of interface
pressure [23], which results in the decrease of dynamic friction
coefcient. More lubrication oil will be squeezed out to the friction
interface from the inner of material under high interface pressure,
so that the oil can cover most of the asperities, which decreases
the mechanical contact [24]. Furthermore, the total energy of
system increases and the braking time is prolonged with the
increase of rotating speed, which signicantly increases the interface temperature. The increasing temperature decreases oil viscosity and then leads to the decline of shear stress produced by oil
lm. Meanwhile, the increase of interface temperature causes
softening of materials, resulting in the decrease of dynamic friction
coefcient. However, the relative size of dynamic friction coefcient is consistent in different interface pressures and at different
rotating speeds. The results are in accordance with the study of
Kim et al. [12], Kim et al. [25] and Berger et al. [26]. They also
found that the friction coefcient decreases with the increase of
interface pressure and rotating speed but the relative size is
consistent. Therefore, the comprehensive evaluation model can
also be applied to the evaluation of the tribological properties in
different interface pressures and at different rotating speeds.

4.1. Application of the model

4.2. Model Checking
In the paper, the paper-based friction materials modied by
carbon nanotubes (CNTs) are taken for example. The paper-based
friction materials containing 0, 2, 4, 8, 12 and 15 wt% of CNTs are
selected as the six projects (P1, P2, P3, P4, P5, P6), as shown in Fig. B3.
4.1.1. Weight calculation
Judgment matrix (a) is obtained by 19 scale method and
expert judgment. The maximum eigenvalue and eigenvector of
judgment matrix (a) are obtained by the software of mat-lab. The
maximum eigenvalue max is 5.0331. The weights of dynamic
friction coefcient 1 , the ratio of dynamic and static friction
coefcient 2 , static friction coefcient 3 , variation coefcient
4 and wear rate 5 are shown in Table C2.
4.1.2. Data normalization
Table C3 shows the experimental values and normalized data of
the ve criteria for the six evaluation projects. The experimental
values are normalized through Eqs. (A.3) and (A.4).

4.2.1. Consistency checking

Consistency checking of judgment matrix is performed through
Eqs. (A.6) and (A.7), where CR is the ratio of coherence; CI is the
index of coherence; RI is the index of average random coherence;
max is the maximum eigenvalue; n is the order of the matrix.
Random consistency index (CR) is 0.0074, which is obviously
less than 0.1294 (RI 1.12). Therefore, the judgment matrix has
very good consistency and can accurately reect the rationality of
the weights of the criteria.
4.2.2. The validity checked by friction torque
The friction torque curve can also reect the tribological
properties of the CNTs modied paper-based friction materials
from another perspective [9,24]. The friction torque curves were
obtained under the same condition with the test of friction
coefcient.
Fig. B6 shows the relationship between engagement time and
friction torque. It is found that P3 (4 wt% of CNTs) has the largest

W. Li et al. / Tribology International 92 (2015) 418424

friction torque, the shortest engagement time and the attest

friction torque curve, followed by P4, P6, P5, P1, P2, implying that
the braking efciency and stability of P3 is better than that of other
samples. So the evaluation result of friction torque is exactly
consistent with that of the fuzzy comprehensive evaluation model.
The results are also in accordance with the ndings of Zhang et al.
[26]. They found that the sample possessing better tribological
properties has atter friction torque curve.

4.2.3. The validity checked by surface structure

The surface structure has a great inuence on the tribological
properties of paper-based friction materials, such as a large
number of micro-pores, asperity summits and grooves, ber
protrusion and exposing [2729]. Thus, the surface structure can
indirectly reect the tribological properties. As shown in Fig. B7,
the bers are dispersed uniformly in resin matrix and some micropores with different sizes are formed. The size and amount of the
micro-pores decrease with the increase of CNTs content. This
porous microstructure enables lubricating oil to reach the contact
zone and keep owing in and out. The lubricating oil can transfer
the frictional heat away and then enhance the tribological properties of the samples. However, the adhesive strength decreases with
the increase of the size and amount of the micro-pores, which will
worsen the tribological properties. Meanwhile, the asperity summits decrease with the increase of CNTs content, which will also
weaken the tribological properties.
It is also found that the surface becomes very dense and
smooth when the CNTs content is more than 4 wt%. And the
surface structure becomes very loose when the CNTs content is
less than 4 wt%. The above analyses are in line with the researches
of Fei et al. [6]. They found that the sample possessing suitable
surface structure (porosity and surface roughness) usually has
good tribological properties. Therefore, P3 (4 wt% of CNTs) has
excellent tribological properties from the view of the surface
structure. The evaluation result of surface structure agrees with
that of the fuzzy comprehensive evaluation model.

4.2.4. The validity checked by temperature

The temperature constantly changes owing to the friction
generated heat, and the highest temperature is varied with the
change of tribological properties during the engagement [3032].
Thus, the temperature can also indirectly evaluate the tribological
properties and the low temperature implies good friction stability
and wear performance. In this paper, the highest temperature of
mating plate was obtained during each engagement under the
same condition with the test of wear rate. It is found that the
temperature rapidly increases and then maintains a constant value
with the increase of engagement cycle, as shown in Fig. B8. It is
also found that P3 has the lowest temperature and best temperature stability, followed by P4, P6, P5, P1, P2, which signies that P3
has the best friction stability and wear performance. Osanai et al.
presented that the temperature of friction surface will be lower
and more stable during the engagement, when the friction
material possesses better tribological properties [33], which is in
line with the evaluation result of the fuzzy comprehensive
evaluation model. Thus, the model is checked to be valid by the
evaluation result of temperature.

Yj

8
>
< 1 x1j 2 x2j  3 x3j  4 x4j  5 x5j
>
:

1

for
for

x1j Z 0:08
x1j o 0:08

421

5. Conclusions
Fuzzy comprehensive evaluation model was established in
order to choose the wet friction materials possessing excellent
tribological properties as the friction component. AHP was introduced into the model and the weight function was established to
make the evaluation more objective and accurate. The paper-based
friction materials containing 0, 2, 4, 8, 12 and 15 wt% of CNTs were
chosen as the six evaluation projects (P1, P2, P3, P4, P5, P6). The main
conclusions from this investigation are as follows:
1. The model exhibits high consistency by the consistency
checking.
2. Evaluation results show that P3 has the biggest comprehensive
evaluation index, indicating the best tribological properties,
followed by P4, P6, P5, P1, and the P2 has the smallest comprehensive evaluation index, indicating the worst tribological
properties, which agrees with the results obtained through
the friction torque, surface structure and temperature.
3. Experiment results show that P3 exhibits the highest dynamic
friction coefcient and the ratio of dynamic and static friction
coefcient, and the lowest static friction coefcient and variation coefcient. The P4 exhibits the smallest wear rate.
4. The model can settle the inconsistency problem of multiple
indexes in qualitative analysis and can greatly reduce the
subjective and lopsided errors. And then, it can be also applied
to other wet friction material.

Acknowledgments
This work was supported by the National Natural Science
Foundation of China (No. 51102196), the Science and Technology
Project of the Young Star of Shaanxi Province (2014KJXX-68), the
Innovation Team Assistance Foundation of Shaanxi Province
(2013KCT-06), the Natural Science Foundation of Shaanxi Province
(No. 2012JQ6010), the Scientic Research Project of Shaanxi
Education Department (14Jk1104), the fund of the State Key
Laboratory of Solidication Processing in NWPU (SKLSP201316)
and Postgraduate Innovative Project of SUST (2014011). Their
supports are gratefully acknowledged.

Appendix A. Mathematical formula

V

A  h
n  12I2

C:V

 100%
m
n
P

Mi

xij

A:1

j1

xij
Mi

and

x2j r 1

and

or

x2j 4 1

or

A:2

xij
i 1; 2; 3; 4; 5
i 1; 2; 3; 4; 5
j 1; 2; 3n

x5j r 6  10  2
x5j 4 6  10  2

A:3

A:4

A:5

422

W. Li et al. / Tribology International 92 (2015) 418424

Fig. B4. Relationship between dynamic friction coefcient and interface pressure.

Fig. B1. Schematic diagram of the samples for tribological properties test.

Fig. B5. Relationship between dynamic friction coefcient and rotating speed.

Fig. B2. Schematic diagram of QM1000-II wet friction performance tester.

(1) Flywheel, (2) speed recorder, (3) clutch, (4) guide for friction plate, (5) sample,
(6) couple plate, (7) hydraulic cylinder, (8) ow meter, (9) oil tank, (10) DC motor,
(11) pyrogenation installation, (12) thermometer, (13) computer, (14) controller,
(15) AC motor and (16) frequency drum.

Fig. B6. Friction torque curves of the samples with different CNTs content.

CI

max  n
n1

A:6

CI
RI

A:7

CR

Appendix B
Fig. B3. Fuzzy comprehensive evaluation model.

See Figs. B1B8

W. Li et al. / Tribology International 92 (2015) 418424

423

Fig. B7. SEM micrographs of samples with different CNTs content. (a) 0%, (b) 2%, (c) 4%, (d) 8%, (e) 12% and (f) 15%.

Table C1
The ingredients ratios (wt%) of the samples.
Raw materials

C1

C2

C3

C4

C5

C6

Carbon ber
Bamboo ber
Phenolic resin
CNTs

53
27
20
0

52
26
20
2

51
25
20
4

48
24
20
8

45
23
20
12

43
22
20
15

Table C2
Weights of the criteria.

Fig. B8. Relationship between the highest temperature of mating plate and
number of engagement.

Appendix C
See Tables C1C4

Weight

0.4030

0.1367

0.0791

0.1367

0.2444

424

W. Li et al. / Tribology International 92 (2015) 418424

Table C3
Experimental values and normalized data of the six projects.
Specimen

x1

x1

x2

x2

x3

x3

x4 (%)

x4

x5

x5

P1 (CNTs 0%)

0.08
54
0.08
65
0.10
19
0.09
92
0.08
93
0.08
77

0.93
16
0.94
36
1.11
16
1.08
22
0.97
42
0.95
67

0.65
70
0.57
91
0.79
36
0.71
32
0.68
12
0.49
49

1.00
59
0.88
66
1.21
50
1.09
19
1.04
29
0.75
77

0.13
00
0.14
73
0.12
84
0.13
91
0.13
11
0.13
84

0.95
79
1.08
53
0.94
61
1.02
49
0.96
60
1.01
98

3.37

1.41
99
1.51
69
0.23
17
0.73
32
0.96
91
1.12
92

11.60  10  5

1.79
52
2.17
13
0.41
48
0.13
77
1.12
51
0.35
59

P2 (CNTs 2%)
P3 (CNTs 4%)
P4 (CNTs 8%)
P5 (CNTs 12%)
P6 (CNTs 15%)

Table C4
Values of comprehensive evaluation index for the six projects.
Value

Y1

Y2

Y3

Y4

Y5

Y6

Yj

 0.1957

 0.3224

0.4062

0.3704

0.0513

0.1671

Appendix D
Matrix a
2
1
3
6 1=3
1
6
6
6
1=4
1=2
a6
6 1=3
1
4
1=2
2

1=2

1=2 7
7
7
1=3 7
7
1=2 7
5
1

References
[1] Yi G, Yan F. Mechanical and tribological properties of phenolic resin-based
friction composites lled with several inorganic llers. Wear 2007;262:1219.
[2] Fujii T, Tohgo K, Wang Y, Shibata J, Shimamura Y, Katayama N, et al. Fatigue
strength of a paper-based friction material under shear-compressive loading.
Strength, Fract Complex 2011;7:18593.
[3] Wang J, Gu M, Songhao B, Ge S. Investigation of the inuence of MoS2 ller on
the tribological properties of carbon ber reinforced nylon 1010 composites.
Wear 2003;255:7749.
[4] Hwang HJ, Jung SL, Cho KH, Kim YJ, Jang H. Tribological performance of brake
friction materials containing carbon nanotubes. Wear 2010;268:51925.
[5] Patnaik A, Kumar M, Satapathy BK, Tomar BS. Performance sensitivity of
hybrid phenolic composites in friction braking: effect of ceramic and aramid
bre combination. Wear 2010;269:8919.
[6] Fei J, Li H-J, Fu Y-W, Qi L-H, Zhang Y-L. Effect of phenolic resin content on
performance of carbon ber reinforced paper-based friction material. Wear
2010;269:53440.
[7] Zhang X, Li K-Z, Li H-J, Fu Y-W, Fei J. Tribological and mechanical properties of
glass ber reinforced paper-based composite friction material. Tribol Int
2014;69:15667.
[8] Holgerson M, Lundberg J. Engagement behaviour of a paper-based wet clutch
part 2: inuence of temperature. Proc Inst Mech Eng Part D: J Automob Eng
1999;213:44955.
[9] Zhang X, Li K, Li H, Fu Y, Fei J. Inuence of compound mineral ber on the
properties of paper-based composite friction material. Proc Inst Mech Eng,
Part J: J Eng Tribol 2013;227:124152.
[10] Kim SJ, Jang H. Friction and wear of friction materials containing two different
phenolic resins reinforced with aramid pulp. Tribol Int 2000;33:47784.
[11] Fei J, Wang HK, Huang JF, Zeng XR, Luo W. Effects of carbon ber length on the
tribological properties of paper-based friction materials. Tribol Int
2014;72:17986.
[12] Kim SH, Jang H. Friction and vibration of brake friction materials reinforced
with chopped glass bers. Tribol Lett 2013;52:3419.

3.60
0.55
1.74
2.30
2.68

14.03  10  5
2.68  10  5
0.89  10  5
7.27  10  5
2.30  10  5

[13] Mustafa A, Abdollah MFB, Ismail N, Amiruddin H, Umehara N. Materials

selection for eco-aware lightweight friction material. Mech Ind
2014;15:27985.
[14] Dadkar N, Tomar BS, Satapathy BK. Evaluation of yash-lled and aramid bre
reinforced hybrid polymer matrix composites (PMC) for friction braking
applications. Mater Des 2009;30:436976.
[15] A Mustafa, MFB Abdollah, Shuhimi FF, Ismail N, Amiruddin H, Umehara N.
Selection and verication of kenaf bres as an alternative friction material
using weighted decision matrix method. Mater Des 2015;67:57782.
[16] Zhao H, Morina A, Neville A, Durham J, Vickerman R. Understanding friction
behavior in automatic transmission uid LVFA test: a new positive curve
parameter to friction coefcient ratio index evaluation. J Tribol
2011;133:021802.
[17] Mortazavi B, Bardon J, Bomm JAS, Ahzi S. A statistical approach for the
evaluation of mechanical properties of silica/epoxy nanocomposite: verication by experiments. Comput Mater Sci 2012;59:10813.
[18] Altuzarra A, Moreno-Jimnez JM, Salvador M. Consensus building in AHPgroup decision making: a Bayesian approach. Oper Res 2010;58:175573.
[19] Valahzaghard MK, Ferdousnejhad M. Ranking insurance rms using AHP and
factor analysis. Manag Sci Lett 2013;3:93742.
[20] Fei J, Li H-J, Huang J-F, Fu Y-W. Study on the friction and wear performance of
carbon fabric/phenolic composites under oil lubricated conditions. Tribol Int
2012;56:307.
[21] Chang Q, Wang K, Zhao H. The friction and wear properties of short carbon/
CNT/PA6 hybrid composites under dry sliding conditions. Int J Polym Mater
2013;62:5403.
[22] Fei J, Li HJ, Fu YW, Qi LH, Zhang YL. Effect of phenolic resin content on
performance of carbon ber reinforced paper-based friction material. Wear
2010;269:53440.
[23] Severin D, Dorsch S. Friction mechanism in industrial brakes. Wear
2001;249:7719.
[24] Fei J, Li H-J, Qi L-H, Fu Y-W, Li X-T. Carbon-ber reinforced paper-based friction
material: study on friction stability as a function of operating variables. J Tribol
2008;130:04160511.
[25] Kim SS, Hwang HJ, Shin MW, Jang H. Friction and vibration of automotive
brake pads containing different abrasive particles. Wear 2011;271:1194202.
[26] Berger EJ, Sadeghi F, Krousgrill CM. Finite element modeling of engagement of
rough and grooved wet clutches. J Tribol 1996;118:13746.
[27] Miyazaki T, Matsumoto T, Yamamoto T. Effect of visco-elastic property on
friction characteristics paper-based friction materials for oil immersed
clutches. J Tribol 1998;120:3938.
[28] Nyman P, Mki R, Olsson R, Ganemi B. Inuence of surface topography on
friction characteristics in wet clutch applications. Wear 2006;261:4652.
[29] El-Tayeb NSM, Yousif BF, Yap TC. An investigation on worn surfaces of chopped
glass bre reinforced polyester through SEM observations. Tribol Int
2008;41:33140.
[30] Gao H, Barber GC. Engagement of a rough, lubricated and grooved disk clutch
ith a porous deformable paper-based friction material. Tribol Trans
2002;45:46470.
[31] Sergienko VP, Tseluev MY. Effect of operation parameters on thermal loading
of wet brake discs. Part 1: problem formulation and methods of study. J Frict
Wear 2012;33:3229.
[32] Larsson R, Marklund P. Wet clutch under limited slip conditions-simplied
testing and simulation. Proc Inst Mech Eng, Part J: J Eng Tribol
2007;221:54551.
[33] Osanai H, Ikeda K, Kato K. Relations between temperature in friction surface
and degradation of friction materials during engaging of wet friction paper.
SAE Technical Paper, Paper number: 900553; 2006.