- Email: [email protected]
Thermoelastic instability of functionally graded coating with arbitrarily varying properties considering contact resistance and frictional heat
Accepted Manuscript
Thermoelastic instability of functionally graded coating with arbitrarily varying properties considering contact resistance and frictional heat Jia-Jia Mao , Liao-Liang Ke , Jie Yang , Sritawat Kitipornchai , Yue-Sheng Wang PII: DOI: Reference:
S0307-904X(16)30618-7 10.1016/j.apm.2016.11.013 APM 11439
To appear in:
Applied Mathematical Modelling
Received date: Revised date: Accepted date:
7 June 2016 5 October 2016 10 November 2016
Please cite this article as: Jia-Jia Mao , Liao-Liang Ke , Jie Yang , Sritawat Kitipornchai , Yue-Sheng Wang , Thermoelastic instability of functionally graded coating with arbitrarily varying properties considering contact resistance and frictional heat, Applied Mathematical Modelling (2016), doi: 10.1016/j.apm.2016.11.013
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Highlights Stability boundaries grow with the increasing gradient index of the FGM coating.
•
Stability boundaries grow with the decreasing thermal contact resistance and friction coefficient.
•
An appropriate gradient type of the FGM coating can adjust the TEI of sliding systems.
AC
CE
PT
ED
M
AN US
CR IP T
•
1
ACCEPTED MANUSCRIPT
Thermoelastic instability of functionally graded coating with arbitrarily varying properties considering contact resistance and
CR IP T
frictional heat
Jia-Jia Mao a,b, Liao-Liang Ke a*, Jie Yang b, Sritawat Kitipornchai c, Yue-Sheng Wang a
Institute of Engineering Mechanics, Beijing Jiaotong University, Beijing 100044, PR China
b
School of Engineering, RMIT University, PO Box 71, Bundoora, VIC 3083, Australia
c
School of Civil Engineering, The University of Queensland, St Lucia, QLD 4072, Australia
AN US
a
Abstract.
Using the homogeneous multi-layered model, this paper studies the thermoelastic instability (TEI) of
M
the functionally graded material (FGM) coating with arbitrary varying properties considering the frictional
ED
heat and thermal contact resistance. A homogeneous half-plane slides on an FGM coated half-plane at the out-of-plane direction under a uniform pressure. The perturbation method and transfer matrix method are
PT
used to deduce the characteristic equation of the TEI problem, which is then solved to obtain the relationship between the critical sliding speed and critical heat flux. The effects of the gradient index and
CE
varying form of material properties of the FGM coating on the stability boundaries are examined. The
AC
results show that FGM coating can adjust the thermoelastic contact stability of sliding systems.
Keywords: Functionally graded materials; Frictional heat; Thermal contact resistance; Thermoelastic instability
2
ACCEPTED MANUSCRIPT *Corresponding author. Tel.: 86-10-51685755; Fax: 86-10-51682094 E-mail address: [email protected] (Liao-Liang Ke) 1. Introduction When two conforming bodies are placed in contact, the contact pressure distribution is sensitive to comparatively small changes in the surface profile. Thermoelastic deformations, though generally small, can therefore have a major effect on systems involving contact. Furthermore, the thermal boundary
CR IP T
conditions at the interface are influenced by the mechanical contact conditions. So, the thermoelastic problems are coupled through the boundary conditions, and as a consequence the steady-state solution may be nonunique and/or unstable [1], which can be divided into two different categories, i.e., frictionally-excited thermoelastic instability (TEI) and static TEI [2].
AN US
In the sliding system, there is a certain critical value of the sliding speed, exceeding which even if a small perturbation in the uniform contact pressure between two sliding half-planes can cause unstable because of the coupled interaction of the frictional heat, thermoelastic distortion and elastic contact. This instability is often called as the frictionally excited TEI [3-5]. Generally, it leads to the establishment of localized high temperature at contact regions known as hot spots [6], which is the directly attributable to
M
the damage and early failure in the sliding system especially for the brake disk in cars and trains. For the
ED
frictionally excited TEI, Burton et al. [5] raised a perturbation method to discuss two flat plates contacting on a straight common edge with sliding parallel to the line of contact. Then, Lee and his co-authors used this method to investigate the frictionally-excited TEI in the automotive disk brake [6] and drum brake
PT
systems [7]. Decuzzi et al. [8] introduced a new two-dimensional analytical model, where metal and
CE
friction disks were replaced by layers with the finite thickness, to consider the frictionally-excited TEI in multi-disk clutches and brakes. Du et al. [9] first gave the numerical implementation of Burton's
AC
perturbation analysis for the TEI problem in the sliding system. Barber and his co-authors determined the critical sliding speed for the TEI for a brake disk [10] and an axisymmetric clutch/brake [11] using the finite element method. Furthermore, some investigators [11-13] presented a series of experimental studies on the hot spots induced from the TEI in the railway and aircraft brakes. Panier et al. [13] classified and explained the thermal gradients appearance on the surface of the railway disc brakes by thermograph measurements with an infrared camera. Recently, Afferrante and his co-authors present a series of excellent works [14-17] on the dynamic thermoelastic instability (DTEI) in the sliding systems, like brakes and clutches. Zelentsov et al. [18] investigated the thermoelastic frictional sliding of a rigid 3
ACCEPTED MANUSCRIPT half-plane against the surface of an elastic coating of another half-plane through quasi-static and dynamic formulations. TEI problems can also be cause by the static conduction of the heat across an interface between two thermoelastic bodies because the extent of the contact area influences the heat conduction problem and in turn depends on the thermoelastic distortion [1, 2]. By introducing a pressure-dependent thermal contact resistance at the contact interface, Barber [19] investigated the static thermoelastic contact showing that
CR IP T
even a small sinusoidal contact pressure perturbation on a nominally uniform pressure between two half-planes could cause the unstable, only if the flux was sufficiently large. Later, Barber and his co-authors analyzed the influence of material properties on the stability criterion for the free interfaces between two half-planes [20], a layer and a half-plane [21] and two layers [22]. Schade and his co-authors
AN US
considered the static TEI of two bonded half planes [23] and a layer bonded to a half plane [24]. Johansson [25] illuminated that there was the pressure-dependent thermal contact resistance at the sliding interface, which may cause an interaction between the above two instability mechanisms. Recently, Ciavarella and his team examined the effect of this interaction on a rod contacting a rigid wall [26], the sliding contact of two half-planes [27, 28] and a rectangular block sliding against a rigid plane [29]. They
ED
thermal contact resistance [29].
M
showed that frictional heating might stabilize the system for some certain forms of pressure-dependent
To satisfy the special function or requirement, functionally graded materials (FGMs) are manufactured by a special spatial gradient in structure and/or composition, which can reduce the
PT
magnitude of residual and thermal stresses, mitigate stress concentration and increase fracture toughness
CE
[30-33]. Many authors studied the contact mechanics of FGMs via theory, experiment and/or numerical simulation. Their results showed that controlling a material property gradient in FGMs could lead to a
AC
significant improvement in the resistance to the contact deformation and damage [34-39]. In the past few years, the frictionally-excited TEI and static TEI of FGMs were investigated by Jang and his co-authors [40, 41] and Ke and his co-authors [2, 42, 43], respectively. Specially, Liu et al. [44] analyzed the dynamic instability of an elastic solid sliding against a functionally graded material coated half-plane without frictional heat. They found that FGMs have the potential application to improve the contact stability of systems. The effects of the frictional heating and pressure-dependent thermal contact resistance on the TEI of FGMs have been extensively studied in isolation, but quite few work was reported to consider the their possible interaction. Only Ke and his co-authors investigated the frictionally excited TEI of an FGM 4
ACCEPTED MANUSCRIPT half-plane sliding against a homogeneous half-plane in the out-of-plane direction with the thermal contact resistance [45]. In this paper, we further investigate the TEI of a homogeneous half-plane sliding out-of-plane on an FGM coated half-plane considered the coupled effect of the frictional heat and thermal contact resistance. The material properties of the FGM coating are assumed to vary arbitrarily along the thickness direction. The homogeneous multi-layered model is used to deal with the arbitrarily varying properties. Using the
CR IP T
perturbation method and transfer matrix method, we can derive the characteristic equation of the TEI problem, which is then solved to obtain the relationship between the critical sliding speed and critical heat flux. The effects of the gradient index and varying form of material properties of the FGM coating on the stability boundaries are discussed in detail.
AN US
Compared with our previous papers [2, 42, 43, 45] on the TEI of FGMs, the new aspects of this paper are: (1) The main concern of these papers was placed on the static TEI of FGMs, which was induced only by the thermal contact resistance. In this paper, we discuss the coupled effect of the frictional heating and thermal contact resistance on the TEI of FGMs. (2) The material properties of FGM coating are allowed to change arbitrarily along the thickness direction. (3) The effect of different types of inhomogeneity on the
M
coupled TEI of FGMs is considered in this paper. (4) Special attention is paid to the effect of graded
ED
variation of the thermal diffusivity coefficient by depth of the coating, which was neglected in previous
2. Formulation
PT
works.
CE
Fig. 1a shows the schematic map of the frictional sliding between an FGM coated half-plane ( y h ) and a homogeneous half-plane ( y h ) with a relative speed V at the out-of-plane direction (i.e. z
AC
direction). The thickness of the FGM coating is h. These two contact bodies are pressed together by a uniform pressure p0. The thermoelastic properties of the upper homogeneous half-plane are expressed as the shear modulus , thermal conductivity coefficient k , thermal expansion coefficient , thermal diffusivity coefficient and Poisson’s radio . The thermoelastic properties of the coating are position-dependent along the thickness direction and follow arbitrary smooth and continuous functions, which are written as the shear modulus y , Poisson’s ratio v y , thermal conductivity k y , 5
ACCEPTED MANUSCRIPT thermal expansion coefficient y and thermal diffusivity coefficient y . The thermoelastic properties of the bottom homogeneous half-plane are equal to those of the FGM coating at y 0 , i.e.,
0 0 , k0 k 0 , 0 0 , 0 0 , 0 0 . Note that many papers studied the TEI of FGMs with the thermoelastic properties varying exponentially [2, 40-45]. In their analysis, the thermal diffusivity coefficient of FGMs treated as a constant
CR IP T
directly for mathematical convenience because it is impossible to obtain the analytical solution for the transient heat conduction equation with graded thermal diffusivity coefficient. To overcome this problem, the homogeneous multi-layered model [39] shown in Fig. 1b is used to simulate the arbitrarily varying material properties of the FGM coating. That is why the homogeneous multi-layered model is employed in the present analysis. The FGM coating is divided into N sub-layers with the equal thickness
AN US
( h j jh / N , j 1, 2,..., N ). In each sub-layer, the thermoelastic properties, j , k j , j , j , and j are constants.
It is assumed that all of the friction-induced heat flow into the two contact bodies. There is a
M
frictional heat fVp caused by the friction coefficient f, normal contact pressure p and sliding speed V during the movement. Since there is a thermal contact resistance at the sliding interface, it is necessary to
ED
define a partitioning parameter of the frictional heat generation [25, 28], which means the proportion of the frictional heat which flows into these two contact bodies.
PT
Supposing the heat flux which enters into the upper homogeneous half-plane is positive. Defining the
CE
temperature of the sub-layer N of the FGM coating and the upper homogeneous half-plane as TN and
T , respectively, there is a temperature drop TN T across the sliding surface. A thermal contact
AC
resistance R p exists and deduces with the pressure at the contacted interface [46]. Thus,
q
TN T fVp , R p
(1)
TN T 1 fVp , R p
(2)
for the upper homogenous half-plane, and
qN
for the sub-layer N of the FGM coating, the subscripts “N” and “ ” respectively represents the sub-layer 6
ACCEPTED MANUSCRIPT N of the FGM coating and the upper homogeneous half-plane. At the interface of y h , the heat flux flows into the upper half-plane q and the sub-layer N qN must satisfy
q qN fVp , y h ,
y h
k
T y
and qN
y h
k hN
y h
TN y
. y h
Using the small perturbation to Eqs. (1) - (3) [1, 45], we have
CR IP T
where q
(3)
q R0 T p Rq 0 fV R0 p0 R , and
T T k hN N fV p 0 , y y
where
AN US
k
R Rp , R dR p / dp , R0 R p0 , T TN T , TN 0 T 0 fVp0 , R0
M
q 0
(4)
(5)
(6)
(7)
ED
with the steady-state values for the thermal contact resistance R0 , the contact pressure p0 , the heat flux
PT
on the upper homogeneous half-plane q 0 , temperature on the sub-layer N of the FGM coating TN 0 and temperature on the upper homogeneous half-plane T 0 . Note that the perturbations of the heat flux for the
CE
upper homogeneous half-plane q , temperature drop T and contact pressure p at the interface
AC
will be settled in the next part.
3. Stability analysis 3.1 Temperature perturbation The temperature perturbations in the upper homogeneous half-plane, FGM coating and bottom homogeneous half-plane can be written as [45]
Tl xl , yl , t fl yl ebt i mxl , l 0,1, 2,..., N , ,
(8)
with i 1 ; the coordinates of the upper homogeneous half-plane, sub-layers of the FGM coating and 7
ACCEPTED MANUSCRIPT bottom homogeneous half-plane xl , yl ; wave number m; complex functions fl yl which can be solved by the transient heat conduction equation and exponential growth rate b which is either real or complex. Referring to Zhang and Barber [20], a positive real b or a complex b with a positive real part could lead the system to be unstable. The temperature perturbations must satisfy the transient heat conduction equations
2T 2T 1 T , 2 2 x y t
x j
2
2T j y j
2
CR IP T
2T j
(9)
1 T j , j 1, 2,..., N , j t
AN US
2T0 2T0 1 T0 , x0 2 y0 2 0 t
(10)
(11)
for the upper homogeneous half-plane, sub-layers 1~N of the FGM coating and bottom homogeneous half-plane, respectively; and the coordinates of them satisfy
x x x j x0 , y y h y j y0 .
(12)
M
Substituting Eqs. (8) and (12) into Eqs. (9), (10) and (11) and considering the condition at the infinity:
ED
T 0 as y and T0 0 as y , we can obtain the temperature perturbation for the upper homogeneous half-plane, sub-layers 1~N of the FGM coating and bottom homogeneous half-plane
PT
T x, y, t C 1e a y hebt imx ,
CE
Tj x, y, t C j1e
a j y
C j 2e
aj y
e
(13) bt imx
,
T0 x, y, t C01ea0 y ebt imx ,
(14) (15)
AC
where C 1 , C j1 , C j 2 and C01 are unknown constants, and
al m2
b
l
, l 0,1, 2,..., N , .
(16)
3.2 Thermoelastic stress and displacement fields Since we employ the homogeneous multi-layered model to the FGM coating, the governing equations of the sub-layers 1~N are the same as those of two homogenous half-planes. They can be written as
8
ACCEPTED MANUSCRIPT 2 2 2uxl u yl 4 l Tl uxl , l 1 x 2 xy l 1 x
(17)
2 2 u yl 2uxl 4 l Tl u yl . l 1 y 2 xy l 1 y
(18)
2
2
The stress field for them is given by
l 1 l u yl l 3 uxl 4ll T
l 1 y
l 1 x l 1
u yl uxl y x
xyl l
.
l,
(19)
CR IP T
yl
(20)
where l 0,1, 2,..., N , ; uxl uxl ( x, y, t ) and u yl u yl ( x, y, t ) are the displacements in the x and y
AN US
directions, respectively; yl and xyl are, respectively, the normal stress and shear stress in the plane of xOy; and
l l (1 l ) , l 3 4 l .
(21)
The temperature perturbation can induce the displacement field among the upper homogeneous
M
half-plane, sub-layers 1~N and bottom homogeneous half-plane. The displacement field is assumed as [2]
uxl x, y, t U xl y ebt imx ,
ED
(22)
u yl x, y, t U yl y ebt imx ,
(23)
PT
where U xl y and U yl y are complex functions of the real variable y.
CE
Substituting Eq. (13) to Eqs. (17) and (18), we obtain
AC
U x ( y)
1 2 4i m 2i m m U x ( y ) U y ( y ) C 1e a y h , 1 1 1
1 4a 2i m U y ( y ) m2U y ( y ) U x ( y ) C 1e a y h . 1 1 1
(24)
(25)
The solutions of Eqs. (24) and (25), which are composed by a homogeneous solution and a particular solution,
ux ( x, y, t ) A 1 A 2 y h e
m y h
A 3e
a y h
e
bt i mx
,
ia u y ( x, y, t ) i A 1 y h A 2 e m y h A 3e a y h ebt i mx . m m 9
(26)
(27)
ACCEPTED MANUSCRIPT Substituting Eq. (14) into Eqs. (17) and (18), we obtain
U xj y
j 1 2 4i m j 2i m a y a y m U xj y U yj y C j1e C j 2e , j 1 j 1 j 1
j
j
j 1 4a j j 2i m C j1e a y C j 2e a y . U yj y m2U yj y U xj y j 1 j 1 j 1 j
j
(28)
(29)
So the displacement field of sub-layers 1~N of the FGM coating can be expressed as a j y
A j 6e
aj y
ebt i mx ,
CR IP T
uxj ( x, y, t ) Aj1 Aj 2 y emy Aj 3 Aj 4 y e my Aj 5e
u yj ( x, y, t ) i Aj1 y j A j 2 e my i A j 3 y j A j 4 e my m m , ia j a y a y + Aj 5e j Aj 6e j ebt i mx m
AN US
(30)
(31)
Similarly, substituting Eq. (15) into Eqs. (17) and (18), we can obtain the displacement field of the bottom homogeneous half-plane [45]
ux 0 ( x, y, t ) A01 A02 y emy A03ea0 y ebt i mx ,
(32)
ia u y 0 ( x, y, t ) i A01 y 0 A02 emy 0 A03e a0 y ebt i mx . m m
M
(33)
ED
The stresses field of the upper homogeneous half-plane, sub-layers 1~N of the FGM coating and bottom homogeneous half-plane can be respectively rewritten as
1 m y h A 2 e m y h mA 3e a y h ebt i mx , 2
(34)
1 2 m y h A 2 e m y h a e a y h A 3 ebt i mx , 2
(35)
PT
y ( x, y, t ) 2i mA 1
CE
AC
xy , ( x, y, t ) 2 mA 1
yj ( x, y, t ) i j 1 j Aj 2 2m Aj1 Aj 2 y e my 1 j Aj 4
a y a y 2m Aj 3 Aj 4 y e my 2m Aj 5 e j Aj 6 e j
ebt i mx
xy , j ( x, y, t ) j 2mAj1 2my 1 j Aj 2 e my 2mAj 3
a y a y 2my 1 j Aj 4 e my 2a j Aj 5 e j 2a j Aj 6 e j ebt i mx
,
1 0 my A02 e my mA03e a0 y ebt i mx , 2
y 0 ( x, y, t ) 2i 0 mA01
,
10
(36)
(37)
(38)
ACCEPTED MANUSCRIPT
1 2 my A02 emy a0e a0 y A03 ebt i mx , 2
xy ,0 ( x, y, t ) 20 mA01
(39)
where
C j1
a
2 j
m2 1
2
4i m
m2 j 1 4i m j
C01
a
2 0
Aj 5 , C j 2
A 3，
a
m2 0 1 4i m 0
2 j
A03 .
m2 j 1 4i m j
Aj 6 ,
(40)
CR IP T
C 1
a
The unknowns A 1 ~ A 3 , Aj1 ~ Aj 6 and A01 ~ A03 could be solved by the boundary conditions. Then,
AN US
q , T and p can be expressed as
q q x, h, t a
a
2
m2 1 4im
A 3ebt i mx ,
M
aN2 m 2 N 1 T TN ( x, h, t ) T ( x, h, t ) 4i m N , 2 2 a m 1 A ebt i mx AN 5e aN h AN 6 e aN h 3 4i m
ED
(41)
PT
1 p y ( x, h, t ) 2i mA 1 A 2 mA 3 ebt i mx . 2
CE
3.3 Boundary conditions 3.3.1 The FGM coating
AC
The displacement field, temperatures, heat flux and stress field are continued at the interface of the
sub-layers, e.g.,
uxj x, h j , t ux, j +1 x, h j , t ，u yj x, h j , t u y , j 1 x, h j , t , Tj x, h j , t T j 1 x, h j , t ，q j x, h j , t q j 1 x, h j , t ,
yj x, h j , t y , j 1 x, h j , t ， xyj x, h j , t xy , j 1 x, h j , t . Then, Eq. (42) can be written in the matrix form 11
(42)
ACCEPTED MANUSCRIPT j h j j j 1 h j j 1 ,
where j Aj1 , Aj 2 , Aj 3 , Aj 4 , Aj 5 , Aj 6
T
(43)
and
j y 1 , 2 , 3 , 4 , 5 , 6 , T
1 = emy，yemy，e my，ye my，e
a j y
，e
aj y
,
j my my j my a j a y a j a y 2 = emy， e , y e ，e ， ye ， e ， j
m
m
m
m
CR IP T
j
a 2j m2 j 1 a j y a 2j m2 j 1 a j y = 0, 0, 0, 0， e ， e , j j 3
a 2j m2 j 1 a j y a 2j m2 j 1 a j y = 0, 0, 0, 0, a j j e , a j j e , j j
=2m e
AN US
4
5 = 2m j emy , j 2my j 1 emy , 2m j e my , j 2my j 1 emy , 2m j e
a j y
, 2m j e
, j 2my j 1 emy , 2m j e my , j 2my j 1 e my , 2 j a j e
a j y
, 2 j a j e
6
j
my
aj y
is the transfer matrix. Finally, we have
M
Thus, j 1 j j , where j j11 h j j h j
, .
aj y
(44)
ED
N N 1N 2 ...11 .
3.3.2 At y 0
PT
In order to maintain the integrity of the FGM coating and bottom homogeneous half-plane, we
CE
assume the boundary conditions at y 0 to be continuous
AC
ux 0 x,0, t ux1 x,0, t ，u y 0 x,0, t u y1 x,0, t , T0 x,0, t T1 x,0, t ，q0 x,0, t q1 x,0, t ,
(45)
y 0 x,0, t y1 x,0, t ， xy 0 x,0, t xy1 x,0, t .
Eq. (45) can be expressed in the matrix form 00 =1 h0 1 , then
1 =1-1 h0 00 , where 0 A01 , A02 , A03 and T
12
(46)
ACCEPTED MANUSCRIPT 0 01 ,02 ,03 , 01 1, 1， 0， 0， 2m0 , 2m0 , T
0， 0 , 0， 0, 0 0 1 , 0 1 0 , m T
2 0
2 2 a02 m2 0 1 a0 a0 m 0 1 1， , , a00 , 2m0 , 2a0 0 . m 0 0 T
CR IP T
3 0
3.3.3 At y h
At the sliding surface ( y h ), the displacements and stresses are continuous [28, 45], i.e.,
AN US
u y x, h, t u yN x, h, t ,
(47)
y x, h, t yN x, h, t ,
(48)
xy x, h, t 0 ,
(49)
M
(50)
where A 1 , A 2 , A 3
= N N
(51)
and
PT
T
ED
Similarity, the matrix form of Eqs. (47)-(50) and (4-5) is
1 ,2 ,3 ,
1 N 1 N * * * 1, 2m , 2m, 0, 2mV * 1 , m Q 2 R V 1 , N 1 N 1 T
CE
AC
1
1 N * 1 1 N * * * , 1 ,1 , 0, 1 1 V , 1 Q 2 R V , 2 N 1 m N 1 T
2
2 2 a 1 N a a m * , 2m , 2a , 0, 2mV 1 , 2 m 1 m N 3
2 2 1 N a m a * * * * m Q 2 R V 1 1 R , 1 m m N T
N N1 , N2 , N3 , N4 , N5 , N6 , T
13
ACCEPTED MANUSCRIPT
N1 emh , h
N
N mh mh e ,e , h m m
mh aN aN h aN aN h e , e , e , m m
N2 2mN emh , N 1 N 2mh emh , 2mN e mh , N 1 N 2mh e mh , 2mN e a h , 2mN ea N
Nh
,
N3 = 0,0,0,0,0,0 , N4 2memh , 2mh N 1 emh , 2me mh , 2mh N 1 e mh , 2aN e a h , 2aN ea N
Nh
,
aN N N 1 aN2 m2 a h aN N N 1 aN2 m2 a h 0, 0, 0, 0, e N , e N , 2 2 m N 1 m N 1
CR IP T
5 N
N 1 aN2 m2 a h N 1 aN2 m2 a h 0, 0, 0, 0, e N , e N , m 1 m 1 N N 6 N
with
TN 0 T 0 1 1 1 1 N , V * =2 fMV , . N R0 mk M 2
AN US
R* mR0 k , Q* 4 RM
Note that R*，Q* and V * are respectively the dimensionless thermal contact resistance, interface
M
temperature drop or heat flux and sliding speed at the interface [28].
ED
From Eqs. (44), (46) and (51), we obtain
0 .
(52)
PT
where N N 1 N 2 ...111 h0 0 . Eq. (52) can be expressed in the dimensionless form as
* *0 ,
(53)
CE
by utilizing the following dimensionless parameters
AC
a0*
a0 a b b 1+ 2 1 z , a* 1 2 1 z , m m 0 0 m m
a*j
aj m
1
b m j 2
1
b z , z 2 , H mh , m j
(54)
(55)
where * and * are the dimensionless forms of and . It is easy to know that * and * are 6 3 matrix, so Eq. (52) can be treated as six homogeneous linearity equations with six unknowns
A01 , A02 , A03 , A 1 , A 2 and A 3 . If and only if the coefficient matrix equals to zero, there are nontrival solutions for these six equations. That is to say, 14
ACCEPTED MANUSCRIPT det * , * 0 .
(56)
Eq. (56) is the characteristic equation of the exponential growth rate z
z b / m . 2
Eq. (56)
includes the interaction of the friction coefficient and the thermal contact resistance, which makes it possible for us to study the coupled effects of them on the stability boundaries meanwhile.
CR IP T
4. Stability criterion It will cause the system to be unstable if the assumed form perturbation has a dimensionless
exponential growth rate z z b / m2
with real and positive or complex with positive real part [20].
To study the stability boundary, we consider two different kinds of the exponential growth z.
AN US
Firstly, z is real and z 0 . The characteristic equation (56) turns to the linear relation between the critical sliding speed V * and critical heat flux Q* . Secondly, z is complex with Re[z] 0 . There are three unknowns in the characteristic equation (56), i.e., the critical sliding speed V * , critical heat flux
Q* and real w. Separating the real and imaginary parts of Eq. (56), we have
(57)
Im det * , * 0 ,
(58)
ED
M
Re det * , * =0 ，
PT
from which we can have the relationship between V * and Q* via the parameter w.
CE
5. Results and discussion
In the sliding systems like brakes and clutches, the frictional heat and thermal contact resistance
AC
always exist at the same time. They will lead to the non-uniform thermal distortions and thermal distribution, which ultimately induce one of the commonest and most dangerous appearances, i.e., the hot spots. It is not accurate enough only considering one of them. In this section, we will focus on the thermoelastic contact instability between a homogeneous half-plane and an FGM coated half-plane with arbitrarily varying properties, taking both the frictional heat and thermal contact resistance into account. In this paper, we mainly study the effects of the thermal diffusivity coefficient , the gradient index, the thickness of the FGM coating H, the thermal contact resistance R* , the friction coefficient f and the heat partition coefficient on the stability boundaries. The FGM coating is made by a mixture of the ceramic 15
ACCEPTED MANUSCRIPT and nodular cast iron. The material of the FGM coating at y 0 is nodular cast iron. The material of the bottom homogeneous half-plane is also taken as the nodular cast iron. The material of the upper homogeneous half-plane is made by the friction material [40]. The properties of the metal, ceramic and friction materials are listed in Table 1 [20, 40]. Since the homogeneous multi-layered model can simulate FGMs with arbitrarily varying properties, we consider FGMs with the power-law form, exponential function, sinusoidal function and cosine function,
the power-law function through the thickness direction, n
1 y
y 0 N 0 , h n
AN US
2 y k y k0 k N k0 , h
CR IP T
respectively. We name the power-law form as Case 1 in which the properties of the FGM coating follow
(59a)
(59b)
n
3 y y 0 N 0 , h
y
n4
M
y 0 N 0 , h
ED
PT CE AC
(59d)
n5
y y 0 N 0 , h
the exponential form
(59c)
(59e)
y 0e y , 1 ln N / 0 / h ,
(60a)
k y k0e2 y , 2 ln kN / k0 / h ,
(60b)
y 0e y , 3 ln N / 0 / h ,
(60c)
y 0e y , 4 ln N / 0 / h ,
(60d)
y 0e y , 5 ln N / 0 / h ,
(60e)
1
3
4
5
as Case 2; the sinusoidal form
y , 2h
y 0 N 0 sin
16
(61a)
ACCEPTED MANUSCRIPT y k y k0 k N k0 sin , 2h
(61b)
y , 2h
(61c)
y , 2h
(61d)
y , 2h
(61e)
y 0 N 0 sin
y 0 N 0 sin as Case 3 and the cosine form
0 N 2
0 N 2
2
y
0 N 2
0 N 2
0 N
M
0 N
ED
y
2
y cos , h
k0 k N k 0 k N y cos , 2 2 h
k y
y
0 N
AN US
y
CR IP T
y 0 N 0 sin
2
0 N 2
(62a)
(62b)
y cos , h
(62c)
y cos , h
(62d)
y cos , h
(62e)
PT
as Case 4, where N , kN , N , N and N are the shear modulus, thermal conductivity coefficient,
CE
thermal expansion coefficient, thermal diffusivity coefficient and Poisson’s radio of the upper surface of
AC
the FGM coating, respectively.
5.1 Convergence and comparison studies Before analyzing the TEI of FGMs, we need to examine the convergence of the homogeneous
multi-layered model. Table 2 shows the effect of the total number of sub-layers (N) on the critical sliding speed V * with H =1.0, R* = 1.0, f = 0.3, Q* 30 and 0.5 . Note that both n 0.1 and stand for the greatly varying material properties of FGM.
n 10
It is observed that the critical sliding speeds
become closer to each other with an increasing N. Obviously, the accurate results can be obtained if the 17
ACCEPTED MANUSCRIPT number of sub-layers N = 8 or 10 for n 1.0 , and N = 12 or 14 for n 0.1 and 10 . Therefore, we choose 12 sub-layers for all examples in the next analysis. It should be pointed out that the extreme gradient slop of FGMs, for example n 0.01 and
n 100 , are beyond the consideration of the
present analysis. Actually, it is hard to fabricate FGMs with the extreme gradient slop. So, the gradient index is limited at the range 0.1 n 10 in our discussion. We choose the stainless and aluminum alloy, respectively, as the material of the upper homogeneous
CR IP T
and bottom homogeneous half-plane. The present problem can be reduced to the static TEI problem [20] if we set the sliding speed and the thickness of the FGM coating as zero. Fig. 2 presents the stability boundaries ( Q* vs R*) for a homogeneous stainless steel half-plane and a homogeneous aluminum alloy half-plane with V * 0.0 . Zhang and Barber’s results are also plotted in Fig. 2 to validate the present
AN US
analysis. Obviously, the present results are in good agreement with Zhang and Barber’s results [20]. If we neglect the sliding speed and bottom homogeneous half-plane, the present coupled TEI problem can be reduced to the static TEI problem between and FGM layer and a homogeneous half-plane reported by Mao et al. [2]. Fig. 3 compares the stability boundaries of the static TEI problem by using the
M
homogeneous multi-layered model and exponential model with H=1.0 and n =0.2. The FGM coating is made by a mixture of the ceramic and nodular cast iron. Again, it can be seen that multi-layered model’s
ED
results agree well with exponential model’s results [2].
PT
5.2 The results for the power-law case
For the sake of convenience, we assume that the gradient indices have the same value n in Eq. (59),
CE
i.e., n1 n2 n3 n4 n5 n . Fig. 4 depicts the effect of the gradient index n on the stability
AC
boundaries with H =1.0, R* = 1.0, f = 0.3 and 0.5 . The stability boundary depends on the real root first and then on the complex root for n = 0.5, 1.0 and 2.0. In the present analysis, we call a special point as the turning point where the instability changes from the real root to the complex root. We can see that the critical sliding speed at the turning point decreases with the decrease of the gradient index n, for example V * 370 for n = 2.0, V * 310 for n = 1.0 and V * 240 for n = 0.5. For the real root instability, the relationship between V * and Q* is linear. However, it is no longer linear for the complex root instability. The same phenomenon is also found by Afferante et al. [28]. For the positive heat 18
ACCEPTED MANUSCRIPT flux, the critical sliding speed decreases with the increase of Q* . But the trend is opposite for the negative heat flux for which the critical sliding speed increases with the increase value of Q* . In other words, the frictional heat can make the system stable in certain situations when considering the pressure-dependent thermal contact resistance [29]. With the increase of the gradient index, the stability area becomes larger. That is because the bigger the gradient index is, the upper sub-layers of the FGM
CR IP T
coating are the ceramic-rich layers, which is more easier to keep stability than the metal-rich layers. Fig. 5 examines the effect of the thickness H on the stability boundaries with n =1.0, R* = 1.0, f = 0.3 and 0.5 . The stability boundaries are all relied on the real roots first and then complex roots for the given thickness H = 1.0, 2.0 or 3.0. Similar to Fig. 3, the relationship between V * and Q* is no longer
AN US
linear for the complex root instability, and the positive heat flux unstabilizes the system, while the negative heat flux stabilizes the system. It is found that the critical speed at the turning point decreases with the increase of H. The dimensionless critical speed is 310 for H = 1.0, 190 for H = 2.0 and 140 for H = 3.0. Note that this conclusion depends on the combinations of coating and substrate materials and the
M
ratios of the material parameters at the bottom and upper surfaces of the coating. If we take other combinations of coating and substrate materials, the conclusion may be opposite from the current one.
H = 1.0, f = 0.3 and
ED
Fig. 6 explores the effect of the thermal contact resistance R* on the stability boundaries with n =1.0,
0.5 . There is still a critical sliding speed V * at the turning point for each curve,
PT
below which the stability boundaries are decided on the real roots, and above which the boundaries depend on the complex roots. Interestingly, the critical sliding speed at the turning points has the same
CE
value 310 for three different R* 0.5, 0.8, 1.0. That is to say, the thermal contact resistance has no
AC
effect on the turning point. However, it has a great effect on the stability area, which decreases with the increase of the thermal contact resistance. The bigger the thermal contact resistance is, the more the frictional heat accumulated at the interface. So the system is much more possible to be unstable. Fig. 7 presents the effect of the friction coefficient f on the stability boundaries with n =1.0, H = 1.0, R*
= 1.0 and
=0.5 . It can be seen that the critical speed V * at the turning point increases with the
decrease of the friction coefficient f, e.g. V * 310 for f = 0.30, V * 370 for f = 0.25 and V * 465 for f = 0.20. The increase of the friction coefficient f leads to the decrease of the stability area. In the 19
ACCEPTED MANUSCRIPT sliding system, the frictional heat is one of the most important factors for the TEI. The bigger the friction coefficient is, the more the heat generated by the friction is. Thus, the system is much more possible to be unstable with the bigger friction coefficient f. Fig. 8 shows the effect of the heat partition coefficient on the stability boundaries with n =1.0, H = 1.0, R* = 1.0 and f = 0.30. It is extremely difficult to confirm the value because it is closely relevant
CR IP T
to the mechanism of the interface. Unless there are available data obtained from the experiments, researchers always select the value within the range of the possible values 0.0~1.0 [28]. Therefore, it is chosen as 0.2, 0.5 and 0.8 in the discussion. It is observed that the turning point is insensitive to the heat partition coefficient. However, the stability area is increasing by the decreasing heat partition coefficient. According to the definition of the heat partition coefficient in section 2, the smaller heat partition
AN US
coefficient indicates that small amount of the frictional heat flows into the upper homogeneous half-plane, but large amount of the frictional heat flows into the FGM coated half-plane.
5.3 Compare the results for different gradient types
M
It is known that the homogeneous multi-layered model has the merit to simulate the arbitrarily
ED
varying material properties of FGMs. By using this model, it is possible to check the effect of different gradient types of FGMs on the stability behavior of system. Indeed, we can select a lot of types of distribution in the analysis. For simplicity, we only select four gradient types of FGMs, i.e, the power-law
PT
function, exponential function, sinusoidal function and cosine function.
CE
Fig. 9 illustrates the stability boundaries for the four different cases with H = 1.0, R* = 1.0, f = 0.3 and
0.5 . For Case 1, the gradient index is takes as n 1.0 . Among the four cases, it is found that Case 1
AC
has the largest stability area and the critical sliding speed at the turning point. The difference of results obtained from different gradient types is great for the complex root instability, while it is relatively small for the real root instability. These results indicate that an appropriate gradient type of the FGM coating can adjust the TEI of systems.
5.4 Discuss the effect of the thermal diffusivity coefficient In previous works concerning the thermoelastic crack problem of FGMs [47, 48], thermoelastic contact of FGMs [36] and TEI of FGMs [40-45], the thermal diffusivity coefficient is always treated as a 20
ACCEPTED MANUSCRIPT constant for the mathematical convenience. In fact, the thermal diffusivity coefficient is inhomogeneous along the thickness direction. Therefore, this treatment of the thermal diffusivity coefficient may lead to the error in the TEI analysis of FGMs. Firstly, we discuss the effect of the thermal diffusivity coefficient on the TEI of FGMs considering the coupled effect of the frictional heat and thermal contact resistance. The properties of the FGM coating follow the power-law function (i.e., case 1). Tables 3 presents the effect of the different form of the
CR IP T
thermal diffusivity coefficient (the graded and the constant ) on the critical sliding speed V * with H =1.0, R* = 1.0, f = 0.3 and 0.5 . Note that 0 and N are the thermal diffusivity coefficient of FGM coating at the bottom surface and upper surface, respectively. “Graded with n =1.0” denotes that the thermal diffusivity coefficient is inhomogeneous along the thickness direction. “ 0 ”,
AN US
“ (0 +N ) 2 ” and “ (0 +N ) 2 ” denote the ways to treat the thermal diffusivity coefficient as constant. The homogeneous multi-layered model is used to solved the TEI of FGMs with the graded or the constant . The values in the bracket denote the relative difference percentages between the solutions of the constant and graded . It is defined as the absolute value of (graded solution -
M
constant solution) / (graded solution). It can be observed that there is no difference between the
ED
graded thermal diffusivity coefficient and the constant for the real root instability. However, for the complex root instability, the different forms of the thermal diffusivity coefficient have a great effect on the
PT
critical sliding speed.
Secondly, we want to examine the error in our previous work about the TEI of FGMs [2, 41-43] with
CE
the thermoelastic parameters varying in the exponential form. In those works, the thermal diffusivity coefficient is treated as a constant by assuming that the gradient indices of the thermal conductivity,
AC
density and specific heat satisfy a certain relation [2, 40-45]. Here, we only consider the effect of the different form of on the critical siding speed V * for the complex root instability since it indeed has no effect on the real root instability behavior. Table 4 tabulates the effect of the thermal diffusivity coefficient on the TEI of FGMs with their properties varying exponentially (i.e., case 2). For the graded , we assume that the gradient indexes have the same value
in Eq. (60), i.e.,
1 2 3 4 5 . For the constant , we have 1 2 3 5 and j 0 or
21
ACCEPTED MANUSCRIPT j N . The bottom surface of the FGM coating is chosen as the nodular cast iron. The gradient index
is respectively selected as 0.1, 0.2, 0.5 and 0.6. It is found that the error is about 0.03%~3.74% between the constant and graded when the gradient index 0.6 . This example indicates that the assumption of constant is reasonable in our previous works [2, 42, 43, 45] when the gradient index
CR IP T
of the exponential function is small.
6. Conclusions
This paper presents the TEI of FGMs considering the coupled effect of the thermal contact resistance and frictional heat in the plane strain state through the perturbation method and the transfer matrix
AN US
method. A homogeneous half-plane slides against an FGM coated half-plane at the out of plane direction. The properties of the FGM coating vary arbitrarily. The homogeneous multi-layered model is used to deal with the arbitrarily varying properties of the FGM coating. The effects of the gradient index, different gradient types and the thermal diffusivity coefficient of the FGM coating on the stability boundaries are discussed. It is found that：
M
(1) For the power law case, the stability boundaries grow with the thermal contact resistance, friction
ED
coefficient, heat partition coefficient and the increasing gradient index of the FGM coating. (2) An appropriate gradient type and thickness of the FGM coating can adjust the TEI of systems. (3) For the power law case, the different form of the thermal diffusivity coefficient of FGM coating has
PT
no effect on the real root instability, but the effect is great for the complex stability behavior. Interestingly,
CE
the results obtained from the average value of the thermal diffusivity coefficient are very close to those obtained from the graded thermal diffusivity coefficient.
AC
(4) For the exponential case, the assumption of the constant thermal diffusivity coefficient is reasonable for the TEI of FGMs when the gradient index is small.
Acknowledgements The work described in this paper is supported by the Fundamental Research Funds for the Central Universities under Grant number 2016YJS106.
References 22
ACCEPTED MANUSCRIPT [1] Barber JR. Thermoelasticity and contact. Journal of Thermal Stresses 1999;22(4-5): 513-25. [2] Mao JJ, Ke LL, Wang YS. Thermoelastic contact instability of a functionally graded layer and a homogeneous half-plane. International Journal of Solids and Structures 2014;51(23): 3962-72. [3] Barber JR. Thermoelastic instabilities in the sliding of comforming solids. In: Proceedings of the Royal Society of London A: Mathematical, Physical and Engineering Sciences. The Royal Society: 1969. p.381-94.
CR IP T
[4] Dow TA, Burton RA. Thermoelastic instability of sliding contact in the absence of wear. Wear 1972;19(3):315-28.
[5] Burton RA, Nerlikar V, Kilaparti SR. Thermoelastic instability in a seal-like configuration. Wear 1973;24(2):177-88.
AN US
[6] Lee K, Barber JR. Frictionally excited thermoelastic instability in automotive disk brakes. Journal of Tribology 1993;115(4):607-14.
[7] Lee K. Frictionally excited thermoelastic instability in automotive drum brakes. Journal of Tribology 2000;122(4):849-55.
[8] Decuzzi P, Ciaverella M, Monno G. Frictionally excited thermoelastic instability in multi-disk clutches
M
and brakes. Journal of Tribology 2001;123(4):865-71.
ED
[9] Du S, Zagrodzki P, Barber JR, Hulbert GM. Finite element analysis of frictionally excited thermoelastic instability. Journal of Thermal Stresses 1997;20(2):185-201. [10] Yi YB, Du S, Barber JR, Fash JW. Effect of geometry on thermoelastic instability in disk brakes and
PT
clutches. Journal of Tribology 1999;121(4):661-6.
CE
[11] Yi YB, Barber JR, Zagrodzki P. Eigenvalue solution of thermoelastic instability problems using Fourier reduction. In: Proceedings of the Royal Society of London A: Mathematical, Physical and
AC
Engineering Sciences. The Royal Society: 2000:2799-821. [12] Audebert N, Barber JR, Zagrodzki P. Buckling of automatic transmission clutch plates due to thermoelastic/plastic residual stresses. Journal of Thermal Stresses 1998;21(3-4):309-26.
[13] Panier S, Dufrenoy P, Bremond P. Infrared characterization of thermal gradients on disc brakes. In: AeroSense 2003:295-302. [14] Afferrante L, Ciavarella M, Barber JR. Sliding thermoelastodynamic instability. Proceedings of the Royal Society of London A: Mathematical, Physical and Engineering Sciences. 2006;462(2071): 2161-2176. 23
ACCEPTED MANUSCRIPT [15] Afferrante L, Ciavarella M. Thermo-elastic dynamic instability (TEDI) in frictional sliding of two elastic half-spaces. Journal of the Mechanics and Physics of Solids. 2007; 55(4):744-764. [16] Afferrante L, Ciavarella M. Thermoelastic dynamic instability (TEDI) in frictional sliding of a half-space against a rigid non-conducting wall. 2007; 74(5): 875-884. [17] Afferrante L, Ciavarella M. Thermo-elastic dynamic instability (TEDI)—a review of recent results. Journal of Engineering Mathematics. 2008; 61:285-300.
CR IP T
[18] Zelentsov VB, Mitrin BI, Aizikovich SM. Dynamic and quasi-static instability of sliding thermoelastic frictional contact. Journal of Friction and Wear. 2016;37(3):213-220.
[19] Barber JR. Stability of thermoelastic contact. In: Proceeding Institution of Mechanical Engineers. International Congress on Tribology 1987:981-86.
AN US
[20] Zhang R, Barber JR. Effect of material properties on the stability of static thermoelastic contact. Journal of Applied Mechanics 1990;57(2):365-9.
[21] Yeo T, Barber JR. Stability of thermoelastic contact of a layer and a half-plane. Journal of Thermal Stresses 1991;14(4):371-88.
Thermal Stresses 1997;20(2):169-84.
M
[22] Li C, Barber JR. Stability of thermoelastic contact of two layers of dissimilar materials. Journal of
ED
[23] Schade DT, Oditt K, Karr DG. Thermoelastic stability of two bonded half planes. Journal of Engineering Mechanics 2000;126(9):981-5. [24] Schade DT, Karr DG. Thermoelastic stability of layer bonded to half plane. Journal of Engineering
PT
Mechanics 2002;128(12):1285-94.
CE
[25] Johansson L. Model and numerical algorithm for sliding contact between two elastic half-planes with frictional heat generation and wear. Wear 1993;160(1):77-93.
AC
[26] Ciavarella M, Johansson L, Afferrante L, Klarbring A, Barber JR. Interaction of thermal contact resistance and frictional heating in thermoelastic instability. International Journal of Solids and Structures 2003;40(21):5583-97.
[27] Afferrante L, Ciavarella M. Frictionally excited thermoelastic instability in the presence of contact resistance. International Journal of Solids and Structures 2004;39(4): 351-7. [28] Afferrante L, Ciavarella M. Instability of thermoelastic contact for two half-planes sliding out-of-plane with contact resistance and frictional heating. Journal of the Mechanics and Physics of Solids 2004;52(7):1527-47. 24
ACCEPTED MANUSCRIPT [29] Ciavarella M, Barber JR. Stability of thermoelastic contact for a rectangular elastic block sliding against a rigid wall. European Journal of Mechanics-A/Solids 2005;24(3):371-6. [30] Suresh S, Mortensen A. Functionally graded metals and metal-ceramic composites: Part 2 Thermomechanical behaviour. International Materials Reviews 1997;42(3):85-116. [31] Krenev LI, Aizikovich SM, Tokovyy YV, Wang YC. Axisymmetric problem on the indentation of a hot circular punch into an arbitrarily nonhomogeneous half-space. International Journal of Solids and
CR IP T
Structures 2015;59:18-28. [32] Chen P, Chen S, Peng Z. Thermo-contact mechanics of a rigid cylindrical punch sliding on a finite graded layer. Acta Mechanica 2012;223(12):2647-65.
[33] Chen P, Chen S. Partial slip contact between a rigid punch with an arbitrary tip-shape and an elastic
AN US
graded solid with a finite thickness. Mechanics of Materials 2013;59:24-35.
[34] Suresh S. Graded materials for resistance to contact deformation and damage. Science 2001;292(5526):2447-51.
[35] Eshraghi I, Dag S, Soltani N. Bending and free vibrations of functionally graded annular and circular micro-plates under thermal loading. Composite Structures 2016 Mar 31;137:196-207.
involving
2011;48(18):2536-48.
frictional
heating.
International
Journal
of
Solids
and
Structures
ED
materials
M
[36] Liu J, Ke LL, Wang YS. Two-dimensional thermoelastic contact problem of functionally graded
[37] Kulchytsky-Zhyhailo R, Bajkowski AS. Axisymmetrical problem of thermoelasticity for halfspace
PT
with gradient coating. International Journal of Mechanical Sciences 2016 Feb 29;106:62-71.
CE
[38] Chen PJ, Chen SH, Yao Y. Nonslipping contact between a mismatch film and a finite-thickness graded substrate. Journal of Applied Mechanics 2016;83(2):021007.
AC
[39] Liu J, Ke LL, Wang YS, Yang J, Alam F. Thermoelastic frictional contact of functionally graded materials with arbitrarily varying properties. International Journal of Mechanical Sciences 2012;63(1):86-98.
[40] Lee SW, Jang YH. Effect of functionally graded material on frictionally excited thermoelastic instability. Wear 2009;266(1):139-46. [41] Lee SW, Jang YH. Frictionally excited thermoelastic instability in a thin layer of functionally graded material sliding between two half-planes. Wear 2009;267(9):1715-22. [42] Mao JJ, Ke LL, Wang YS. Thermoelastic instability of functionally graded materials in frictionless 25
ACCEPTED MANUSCRIPT contact. Acta Mechanica 2015; 226: 2295-311. [43]Mao JJ, Ke LL, Wang YS. Thermoelastic instability of a functionally graded layer interacting with a homogeneous layer. International Journal of Mechanical Sciences 2015; 99:218-27. [44]Liu J, Wang YS, Ke LL, Yang J, Alam F. Dynamic instability of an elastic solid sliding against a functionally graded material coated half-plane. International Journal of Mechanical Sciences 2014;89:323-31.
materials
sliding
out-of-plane
with
contact
2016;83(2):021010.
resistance.
CR IP T
[45] Mao JJ, Ke LL, Wang YS, Liu J. Frictionally excited thermoelastic instability of functionally graded Journal
of
Applied
Mechanics
[46] Li C. Thermoelastic contact stability analysis, Ph.D. thesis. University of Michigan, Michigan: 1998.
AN US
[47] Jin ZH, Noda N. Transient thermal stress intensity factors for a crack in semi-infinite plate of a functionally gradient material. International Journal of Solids and Structures 1994; 31:203-18. [48] Shodja HM, Ghahremaninejad A. An FGM coated elastic solid under thermomechanical loading: a
AC
CE
PT
ED
M
two dimensional linear elastic approach. Surface and Coatings Technology 2006; 200(12):4050-64.
26
ACCEPTED MANUSCRIPT Table captions
Table 1 Thermoelastic properties of selected materials Table 2 The effect of the total number of multi-layers (N) on the critical sliding speed V* for different gradient types Table 3 The effect of the thermal diffusivity coefficient on the critical sliding speed V*: Power-law
CR IP T
function (Case 1) Table 4 The effect of the thermal diffusivity coefficient on the critical sliding speed V* with Q* = -180:
AC
CE
PT
ED
M
AN US
exponential function (Case 2)
27
ACCEPTED MANUSCRIPT Figure captions
Fig. 1 Schematic map of the thermoelastic contact problem for a homogeneous half-plane sliding out of an FGM coated half-plane (a) and the homogeneous multi-layered model (b). Fig. 2 The stability boundaries for a homogeneous stainless half-plane and a homogeneous aluminum alloy half-plane with V* = 0.0 with R* = 1.0, f = 0.3 and 0.5 .
CR IP T
Fig. 3 The stability boundaries of the static TEI problem with H = 1.0 and n 0.2 .
Fig. 4 The effect of gradient index n on the stability boundaries with H =1.0, R* = 1.0, f = 0.3 and
0.5 .
Fig. 5 The effect of the thickness H on the stability boundaries with n =1.0, R* = 1.0, f = 0.3 and 0.5 .
AN US
Fig. 6 The effect of thermal contact resistance R* on the stability boundaries with n =1.0, H = 1.0, f = 0.3 and 0.5 .
Fig. 7 The effect of the friction coefficient f on the stability boundaries with n = 1.0, H = 1.0, R* = 1.0
M
and 0.5 .
Fig. 8 The effect of heat partition coefficient on the stability boundaries with n =1.0, H = 1.0, R* = 1.0
ED
and f 0.3 .
AC
CE
PT
Fig. 9 The stability boundaries for the four different cases with H = 1.0, R* = 1.0, f = 0.3 and =0.5 .
28
ACCEPTED MANUSCRIPT
Table 1 Thermoelastic properties of selected materials Friction
Nodular
Ceramic
Aluminium
Stainless
material
cast iron
alloy
steel
0.3
168
151
72
14.0
13.7
12.0
22.0
k (W/moC)
0.241
48.9
3.0
( mm2/s)
0.177
16.05
1.15
0.12
0.31
SiC
Properties
CR IP T
6
0.30
AC
CE
PT
ED
M
o C1 10
200
400
14.0
4.4
173.0
21.0
110.0
67.16
5.93
35.48
0.30
0.16
AN US
E (GPa)
29
sintered
0.32
ACCEPTED MANUSCRIPT
Table 2 The effect of the total number of multi-layers (N) on the critical sliding speed V* for different gradient types
6
8
10
12
14
16
n = 0.1
93.07
94.44
95.02
95.35
95.55
95.70
95.80
n = 1.0
100.06
102.86
103.93
104.47
104.80
105.02
105.17
n = 10.0
101.81
105.29
106.65
108.07
108.27
AN US
4
CR IP T
Total number of multi-layers (N) n
AC
CE
PT
ED
M
107.36
30
107.79
ACCEPTED MANUSCRIPT
Table 3 The effect of the thermal diffusivity coefficient on the critical sliding speed V*: Power-law function (Case 1)
2
-30
6.40
104.80
Graded with
0
6.40
N
6.40
6.40 Real
347.11
398.96
320.57
333.25
(7.65%)
(16.47%)
343.44
390.17
(1.06%)
(2.20%)
379.29
491.42
(9.27%)
(23.18%)
Complex
Complex
104.80
104.80 Real
AC
CE
PT
ED
Instability Types
-160
104.80
M
(0 +N ) 2
-120
AN US
n =1.0
CR IP T
Q*
31
ACCEPTED MANUSCRIPT
Table 4 The effect of the thermal diffusivity coefficient on the critical sliding speed V* with Q* = -180:
0.2
0.5
0.6
Graded
592.72
585.30
540.55
518.40
0
592.87
586.47
550.12
533.27
(0.03%)
(0.20%)
(1.77%)
(2.87%)
AN US
0.1
592.54
N
CR IP T
exponential function (Case 2)
(0.03%) Instability Types
583.98
527.92
499.00
(0.23%)
(2.33%)
(3.74%)
AC
CE
PT
ED
M
Complex
32
ACCEPTED MANUSCRIPT
AN US
CR IP T
Interface
AC
CE
PT
ED
M
(a)
(b)
Fig. 1 Schematic map of the thermoelastic contact problem for a homogeneous half-plane sliding out of an FGM coated half-plane (a) and the homogeneous multi-layered model (b).
33
ACCEPTED MANUSCRIPT
30
0
CR IP T
Q
*
-30
Present results: Real root Complex root
-60
-90
-120 0.0
0.2
0.4
AN US
Zhang and Barber's results: Real root Complex root
*
0.6
0.8
1.0
1/(1+R )
M
Fig. 2 The stability boundaries for a homogeneous stainless half-plane and a homogeneous aluminum
AC
CE
PT
ED
alloy half-plane with V* = 0.0 with R* = 1.0, f = 0.3 and 0.5 .
34
ACCEPTED MANUSCRIPT
120
Exponential model [2] Multi-layered model
CR IP T
105
Q
*
90
75
AN US
60
45 0.00
0.15
0.30
0.45
0.60
*
M
1/(1+R )
AC
CE
PT
ED
Fig. 3 The stability boundaries of the static TEI problem with H = 1.0 and n 0.2 .
35
ACCEPTED MANUSCRIPT
20 0
CR IP T
-20 -40
*
Q
-80 -100 -120 -140
Instability
Stability
-160 0
50
100
150
AN US
Real roots: n = 0.5 n = 1.0 n = 5.0 Complex roots n = 0.5 n = 1.0 n = 5.0
-60
200
250
M
V
300
350
400
450
*
0.5 .
AC
CE
PT
ED
Fig. 4 The effect of the gradient index n on the stability boundaries with H =1.0, R* = 1.0, f = 0.3 and
36
ACCEPTED MANUSCRIPT
20 0
Real roots: H = 1.0 H = 2.0 H = 3.0
CR IP T
-20
Instability
-40
-80
Complex roots: H = 1.0 H = 2.0 H = 3.0
-100 -120 -140
Stability
-160 0
50
100
AN US
Q
*
-60
150
200
250
300
350
400
*
M
V
AC
CE
PT
ED
Fig. 5 The effect of the thickness H on the stability boundaries with n =1.0, R* = 1.0, f = 0.3 and 0.5 .
37
ACCEPTED MANUSCRIPT
20 0 -20 -40
CR IP T
Instablity Real roots: * R = 0.5 * R = 0.8 * R = 1.0
Q
*
-60 -80 -100
Complex roots: * R = 0.5 * R = 0.8 * R = 1.0
-140
Stablity
AN US
-120
-160 0
50
100
150
200
250
300
350
400
*
M
V
Fig. 6 The effect of the thermal contact resistance R* on the stability boundaries with n =1.0, H = 1.0, f =
AC
CE
PT
ED
0.3 and 0.5 .
38
ACCEPTED MANUSCRIPT
20 0 -20
Complex roots: f = 0.30 f = 0.25 f = 0.20
CR IP T
Instability
-40
Real roots: f = 0.20 f = 0.25 f = 0.30
-80 -100 -120
AN US
Q
*
-60
Stability
-140 -160 0
50
100
150
200
250
300
350
400
450
500
*
M
V
0.5 .
AC
CE
PT
ED
Fig. 7 The effect of the friction coefficient f on the stability boundaries with n = 1.0, H = 1.0, R* = 1.0 and
39
ACCEPTED MANUSCRIPT
30 0
Real roots:
*
-60
Q
CR IP T
Instability
-30
-90 -120
Complex roots:
-180
Stability
AN US
-150
-210 0
50
100
150
200
250
300
350
400
*
M
V
ED
Fig. 8 The effect of heat partition coefficient on the stability boundaries with n =1.0, H = 1.0,
AC
CE
PT
R* 1.0 and f = 0.3.
40
ACCEPTED MANUSCRIPT
20 0
Real roots: Case 1 (n = 1.0) Case 2 Case 3 Case 4
CR IP T
Instability
-20 -40
Stability
-80
Complex roots: Case 1 (n = 1.0) Case 2 Case 3 Case 4
-100 -120 -140 -160 0
50
100
150
AN US
Q
*
-60
200
250
300
350
400
*
M
V
AC
CE
PT
ED
Fig. 9 The stability boundaries for the four different cases with H = 1.0, R* = 1.0, f = 0.3 and =0.5 .
41
Thermoelastic instability of functionally graded coating with arbitrarily varying properties considering contact resistance and frictional heat Jia-Jia Mao , Liao-Liang Ke , Jie Yang , Sritawat Kitipornchai , Yue-Sheng Wang PII: DOI: Reference:
S0307-904X(16)30618-7 10.1016/j.apm.2016.11.013 APM 11439
To appear in:
Applied Mathematical Modelling
Received date: Revised date: Accepted date:
7 June 2016 5 October 2016 10 November 2016
Please cite this article as: Jia-Jia Mao , Liao-Liang Ke , Jie Yang , Sritawat Kitipornchai , Yue-Sheng Wang , Thermoelastic instability of functionally graded coating with arbitrarily varying properties considering contact resistance and frictional heat, Applied Mathematical Modelling (2016), doi: 10.1016/j.apm.2016.11.013
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Highlights Stability boundaries grow with the increasing gradient index of the FGM coating.
•
Stability boundaries grow with the decreasing thermal contact resistance and friction coefficient.
•
An appropriate gradient type of the FGM coating can adjust the TEI of sliding systems.
AC
CE
PT
ED
M
AN US
CR IP T
•
1
ACCEPTED MANUSCRIPT
Thermoelastic instability of functionally graded coating with arbitrarily varying properties considering contact resistance and
CR IP T
frictional heat
Jia-Jia Mao a,b, Liao-Liang Ke a*, Jie Yang b, Sritawat Kitipornchai c, Yue-Sheng Wang a
Institute of Engineering Mechanics, Beijing Jiaotong University, Beijing 100044, PR China
b
School of Engineering, RMIT University, PO Box 71, Bundoora, VIC 3083, Australia
c
School of Civil Engineering, The University of Queensland, St Lucia, QLD 4072, Australia
AN US
a
Abstract.
Using the homogeneous multi-layered model, this paper studies the thermoelastic instability (TEI) of
M
the functionally graded material (FGM) coating with arbitrary varying properties considering the frictional
ED
heat and thermal contact resistance. A homogeneous half-plane slides on an FGM coated half-plane at the out-of-plane direction under a uniform pressure. The perturbation method and transfer matrix method are
PT
used to deduce the characteristic equation of the TEI problem, which is then solved to obtain the relationship between the critical sliding speed and critical heat flux. The effects of the gradient index and
CE
varying form of material properties of the FGM coating on the stability boundaries are examined. The
AC
results show that FGM coating can adjust the thermoelastic contact stability of sliding systems.
Keywords: Functionally graded materials; Frictional heat; Thermal contact resistance; Thermoelastic instability
2
ACCEPTED MANUSCRIPT *Corresponding author. Tel.: 86-10-51685755; Fax: 86-10-51682094 E-mail address: [email protected] (Liao-Liang Ke) 1. Introduction When two conforming bodies are placed in contact, the contact pressure distribution is sensitive to comparatively small changes in the surface profile. Thermoelastic deformations, though generally small, can therefore have a major effect on systems involving contact. Furthermore, the thermal boundary
CR IP T
conditions at the interface are influenced by the mechanical contact conditions. So, the thermoelastic problems are coupled through the boundary conditions, and as a consequence the steady-state solution may be nonunique and/or unstable [1], which can be divided into two different categories, i.e., frictionally-excited thermoelastic instability (TEI) and static TEI [2].
AN US
In the sliding system, there is a certain critical value of the sliding speed, exceeding which even if a small perturbation in the uniform contact pressure between two sliding half-planes can cause unstable because of the coupled interaction of the frictional heat, thermoelastic distortion and elastic contact. This instability is often called as the frictionally excited TEI [3-5]. Generally, it leads to the establishment of localized high temperature at contact regions known as hot spots [6], which is the directly attributable to
M
the damage and early failure in the sliding system especially for the brake disk in cars and trains. For the
ED
frictionally excited TEI, Burton et al. [5] raised a perturbation method to discuss two flat plates contacting on a straight common edge with sliding parallel to the line of contact. Then, Lee and his co-authors used this method to investigate the frictionally-excited TEI in the automotive disk brake [6] and drum brake
PT
systems [7]. Decuzzi et al. [8] introduced a new two-dimensional analytical model, where metal and
CE
friction disks were replaced by layers with the finite thickness, to consider the frictionally-excited TEI in multi-disk clutches and brakes. Du et al. [9] first gave the numerical implementation of Burton's
AC
perturbation analysis for the TEI problem in the sliding system. Barber and his co-authors determined the critical sliding speed for the TEI for a brake disk [10] and an axisymmetric clutch/brake [11] using the finite element method. Furthermore, some investigators [11-13] presented a series of experimental studies on the hot spots induced from the TEI in the railway and aircraft brakes. Panier et al. [13] classified and explained the thermal gradients appearance on the surface of the railway disc brakes by thermograph measurements with an infrared camera. Recently, Afferrante and his co-authors present a series of excellent works [14-17] on the dynamic thermoelastic instability (DTEI) in the sliding systems, like brakes and clutches. Zelentsov et al. [18] investigated the thermoelastic frictional sliding of a rigid 3
ACCEPTED MANUSCRIPT half-plane against the surface of an elastic coating of another half-plane through quasi-static and dynamic formulations. TEI problems can also be cause by the static conduction of the heat across an interface between two thermoelastic bodies because the extent of the contact area influences the heat conduction problem and in turn depends on the thermoelastic distortion [1, 2]. By introducing a pressure-dependent thermal contact resistance at the contact interface, Barber [19] investigated the static thermoelastic contact showing that
CR IP T
even a small sinusoidal contact pressure perturbation on a nominally uniform pressure between two half-planes could cause the unstable, only if the flux was sufficiently large. Later, Barber and his co-authors analyzed the influence of material properties on the stability criterion for the free interfaces between two half-planes [20], a layer and a half-plane [21] and two layers [22]. Schade and his co-authors
AN US
considered the static TEI of two bonded half planes [23] and a layer bonded to a half plane [24]. Johansson [25] illuminated that there was the pressure-dependent thermal contact resistance at the sliding interface, which may cause an interaction between the above two instability mechanisms. Recently, Ciavarella and his team examined the effect of this interaction on a rod contacting a rigid wall [26], the sliding contact of two half-planes [27, 28] and a rectangular block sliding against a rigid plane [29]. They
ED
thermal contact resistance [29].
M
showed that frictional heating might stabilize the system for some certain forms of pressure-dependent
To satisfy the special function or requirement, functionally graded materials (FGMs) are manufactured by a special spatial gradient in structure and/or composition, which can reduce the
PT
magnitude of residual and thermal stresses, mitigate stress concentration and increase fracture toughness
CE
[30-33]. Many authors studied the contact mechanics of FGMs via theory, experiment and/or numerical simulation. Their results showed that controlling a material property gradient in FGMs could lead to a
AC
significant improvement in the resistance to the contact deformation and damage [34-39]. In the past few years, the frictionally-excited TEI and static TEI of FGMs were investigated by Jang and his co-authors [40, 41] and Ke and his co-authors [2, 42, 43], respectively. Specially, Liu et al. [44] analyzed the dynamic instability of an elastic solid sliding against a functionally graded material coated half-plane without frictional heat. They found that FGMs have the potential application to improve the contact stability of systems. The effects of the frictional heating and pressure-dependent thermal contact resistance on the TEI of FGMs have been extensively studied in isolation, but quite few work was reported to consider the their possible interaction. Only Ke and his co-authors investigated the frictionally excited TEI of an FGM 4
ACCEPTED MANUSCRIPT half-plane sliding against a homogeneous half-plane in the out-of-plane direction with the thermal contact resistance [45]. In this paper, we further investigate the TEI of a homogeneous half-plane sliding out-of-plane on an FGM coated half-plane considered the coupled effect of the frictional heat and thermal contact resistance. The material properties of the FGM coating are assumed to vary arbitrarily along the thickness direction. The homogeneous multi-layered model is used to deal with the arbitrarily varying properties. Using the
CR IP T
perturbation method and transfer matrix method, we can derive the characteristic equation of the TEI problem, which is then solved to obtain the relationship between the critical sliding speed and critical heat flux. The effects of the gradient index and varying form of material properties of the FGM coating on the stability boundaries are discussed in detail.
AN US
Compared with our previous papers [2, 42, 43, 45] on the TEI of FGMs, the new aspects of this paper are: (1) The main concern of these papers was placed on the static TEI of FGMs, which was induced only by the thermal contact resistance. In this paper, we discuss the coupled effect of the frictional heating and thermal contact resistance on the TEI of FGMs. (2) The material properties of FGM coating are allowed to change arbitrarily along the thickness direction. (3) The effect of different types of inhomogeneity on the
M
coupled TEI of FGMs is considered in this paper. (4) Special attention is paid to the effect of graded
ED
variation of the thermal diffusivity coefficient by depth of the coating, which was neglected in previous
2. Formulation
PT
works.
CE
Fig. 1a shows the schematic map of the frictional sliding between an FGM coated half-plane ( y h ) and a homogeneous half-plane ( y h ) with a relative speed V at the out-of-plane direction (i.e. z
AC
direction). The thickness of the FGM coating is h. These two contact bodies are pressed together by a uniform pressure p0. The thermoelastic properties of the upper homogeneous half-plane are expressed as the shear modulus , thermal conductivity coefficient k , thermal expansion coefficient , thermal diffusivity coefficient and Poisson’s radio . The thermoelastic properties of the coating are position-dependent along the thickness direction and follow arbitrary smooth and continuous functions, which are written as the shear modulus y , Poisson’s ratio v y , thermal conductivity k y , 5
ACCEPTED MANUSCRIPT thermal expansion coefficient y and thermal diffusivity coefficient y . The thermoelastic properties of the bottom homogeneous half-plane are equal to those of the FGM coating at y 0 , i.e.,
0 0 , k0 k 0 , 0 0 , 0 0 , 0 0 . Note that many papers studied the TEI of FGMs with the thermoelastic properties varying exponentially [2, 40-45]. In their analysis, the thermal diffusivity coefficient of FGMs treated as a constant
CR IP T
directly for mathematical convenience because it is impossible to obtain the analytical solution for the transient heat conduction equation with graded thermal diffusivity coefficient. To overcome this problem, the homogeneous multi-layered model [39] shown in Fig. 1b is used to simulate the arbitrarily varying material properties of the FGM coating. That is why the homogeneous multi-layered model is employed in the present analysis. The FGM coating is divided into N sub-layers with the equal thickness
AN US
( h j jh / N , j 1, 2,..., N ). In each sub-layer, the thermoelastic properties, j , k j , j , j , and j are constants.
It is assumed that all of the friction-induced heat flow into the two contact bodies. There is a
M
frictional heat fVp caused by the friction coefficient f, normal contact pressure p and sliding speed V during the movement. Since there is a thermal contact resistance at the sliding interface, it is necessary to
ED
define a partitioning parameter of the frictional heat generation [25, 28], which means the proportion of the frictional heat which flows into these two contact bodies.
PT
Supposing the heat flux which enters into the upper homogeneous half-plane is positive. Defining the
CE
temperature of the sub-layer N of the FGM coating and the upper homogeneous half-plane as TN and
T , respectively, there is a temperature drop TN T across the sliding surface. A thermal contact
AC
resistance R p exists and deduces with the pressure at the contacted interface [46]. Thus,
q
TN T fVp , R p
(1)
TN T 1 fVp , R p
(2)
for the upper homogenous half-plane, and
qN
for the sub-layer N of the FGM coating, the subscripts “N” and “ ” respectively represents the sub-layer 6
ACCEPTED MANUSCRIPT N of the FGM coating and the upper homogeneous half-plane. At the interface of y h , the heat flux flows into the upper half-plane q and the sub-layer N qN must satisfy
q qN fVp , y h ,
y h
k
T y
and qN
y h
k hN
y h
TN y
. y h
Using the small perturbation to Eqs. (1) - (3) [1, 45], we have
CR IP T
where q
(3)
q R0 T p Rq 0 fV R0 p0 R , and
T T k hN N fV p 0 , y y
where
AN US
k
R Rp , R dR p / dp , R0 R p0 , T TN T , TN 0 T 0 fVp0 , R0
M
q 0
(4)
(5)
(6)
(7)
ED
with the steady-state values for the thermal contact resistance R0 , the contact pressure p0 , the heat flux
PT
on the upper homogeneous half-plane q 0 , temperature on the sub-layer N of the FGM coating TN 0 and temperature on the upper homogeneous half-plane T 0 . Note that the perturbations of the heat flux for the
CE
upper homogeneous half-plane q , temperature drop T and contact pressure p at the interface
AC
will be settled in the next part.
3. Stability analysis 3.1 Temperature perturbation The temperature perturbations in the upper homogeneous half-plane, FGM coating and bottom homogeneous half-plane can be written as [45]
Tl xl , yl , t fl yl ebt i mxl , l 0,1, 2,..., N , ,
(8)
with i 1 ; the coordinates of the upper homogeneous half-plane, sub-layers of the FGM coating and 7
ACCEPTED MANUSCRIPT bottom homogeneous half-plane xl , yl ; wave number m; complex functions fl yl which can be solved by the transient heat conduction equation and exponential growth rate b which is either real or complex. Referring to Zhang and Barber [20], a positive real b or a complex b with a positive real part could lead the system to be unstable. The temperature perturbations must satisfy the transient heat conduction equations
2T 2T 1 T , 2 2 x y t
x j
2
2T j y j
2
CR IP T
2T j
(9)
1 T j , j 1, 2,..., N , j t
AN US
2T0 2T0 1 T0 , x0 2 y0 2 0 t
(10)
(11)
for the upper homogeneous half-plane, sub-layers 1~N of the FGM coating and bottom homogeneous half-plane, respectively; and the coordinates of them satisfy
x x x j x0 , y y h y j y0 .
(12)
M
Substituting Eqs. (8) and (12) into Eqs. (9), (10) and (11) and considering the condition at the infinity:
ED
T 0 as y and T0 0 as y , we can obtain the temperature perturbation for the upper homogeneous half-plane, sub-layers 1~N of the FGM coating and bottom homogeneous half-plane
PT
T x, y, t C 1e a y hebt imx ,
CE
Tj x, y, t C j1e
a j y
C j 2e
aj y
e
(13) bt imx
,
T0 x, y, t C01ea0 y ebt imx ,
(14) (15)
AC
where C 1 , C j1 , C j 2 and C01 are unknown constants, and
al m2
b
l
, l 0,1, 2,..., N , .
(16)
3.2 Thermoelastic stress and displacement fields Since we employ the homogeneous multi-layered model to the FGM coating, the governing equations of the sub-layers 1~N are the same as those of two homogenous half-planes. They can be written as
8
ACCEPTED MANUSCRIPT 2 2 2uxl u yl 4 l Tl uxl , l 1 x 2 xy l 1 x
(17)
2 2 u yl 2uxl 4 l Tl u yl . l 1 y 2 xy l 1 y
(18)
2
2
The stress field for them is given by
l 1 l u yl l 3 uxl 4ll T
l 1 y
l 1 x l 1
u yl uxl y x
xyl l
.
l,
(19)
CR IP T
yl
(20)
where l 0,1, 2,..., N , ; uxl uxl ( x, y, t ) and u yl u yl ( x, y, t ) are the displacements in the x and y
AN US
directions, respectively; yl and xyl are, respectively, the normal stress and shear stress in the plane of xOy; and
l l (1 l ) , l 3 4 l .
(21)
The temperature perturbation can induce the displacement field among the upper homogeneous
M
half-plane, sub-layers 1~N and bottom homogeneous half-plane. The displacement field is assumed as [2]
uxl x, y, t U xl y ebt imx ,
ED
(22)
u yl x, y, t U yl y ebt imx ,
(23)
PT
where U xl y and U yl y are complex functions of the real variable y.
CE
Substituting Eq. (13) to Eqs. (17) and (18), we obtain
AC
U x ( y)
1 2 4i m 2i m m U x ( y ) U y ( y ) C 1e a y h , 1 1 1
1 4a 2i m U y ( y ) m2U y ( y ) U x ( y ) C 1e a y h . 1 1 1
(24)
(25)
The solutions of Eqs. (24) and (25), which are composed by a homogeneous solution and a particular solution,
ux ( x, y, t ) A 1 A 2 y h e
m y h
A 3e
a y h
e
bt i mx
,
ia u y ( x, y, t ) i A 1 y h A 2 e m y h A 3e a y h ebt i mx . m m 9
(26)
(27)
ACCEPTED MANUSCRIPT Substituting Eq. (14) into Eqs. (17) and (18), we obtain
U xj y
j 1 2 4i m j 2i m a y a y m U xj y U yj y C j1e C j 2e , j 1 j 1 j 1
j
j
j 1 4a j j 2i m C j1e a y C j 2e a y . U yj y m2U yj y U xj y j 1 j 1 j 1 j
j
(28)
(29)
So the displacement field of sub-layers 1~N of the FGM coating can be expressed as a j y
A j 6e
aj y
ebt i mx ,
CR IP T
uxj ( x, y, t ) Aj1 Aj 2 y emy Aj 3 Aj 4 y e my Aj 5e
u yj ( x, y, t ) i Aj1 y j A j 2 e my i A j 3 y j A j 4 e my m m , ia j a y a y + Aj 5e j Aj 6e j ebt i mx m
AN US
(30)
(31)
Similarly, substituting Eq. (15) into Eqs. (17) and (18), we can obtain the displacement field of the bottom homogeneous half-plane [45]
ux 0 ( x, y, t ) A01 A02 y emy A03ea0 y ebt i mx ,
(32)
ia u y 0 ( x, y, t ) i A01 y 0 A02 emy 0 A03e a0 y ebt i mx . m m
M
(33)
ED
The stresses field of the upper homogeneous half-plane, sub-layers 1~N of the FGM coating and bottom homogeneous half-plane can be respectively rewritten as
1 m y h A 2 e m y h mA 3e a y h ebt i mx , 2
(34)
1 2 m y h A 2 e m y h a e a y h A 3 ebt i mx , 2
(35)
PT
y ( x, y, t ) 2i mA 1
CE
AC
xy , ( x, y, t ) 2 mA 1
yj ( x, y, t ) i j 1 j Aj 2 2m Aj1 Aj 2 y e my 1 j Aj 4
a y a y 2m Aj 3 Aj 4 y e my 2m Aj 5 e j Aj 6 e j
ebt i mx
xy , j ( x, y, t ) j 2mAj1 2my 1 j Aj 2 e my 2mAj 3
a y a y 2my 1 j Aj 4 e my 2a j Aj 5 e j 2a j Aj 6 e j ebt i mx
,
1 0 my A02 e my mA03e a0 y ebt i mx , 2
y 0 ( x, y, t ) 2i 0 mA01
,
10
(36)
(37)
(38)
ACCEPTED MANUSCRIPT
1 2 my A02 emy a0e a0 y A03 ebt i mx , 2
xy ,0 ( x, y, t ) 20 mA01
(39)
where
C j1
a
2 j
m2 1
2
4i m
m2 j 1 4i m j
C01
a
2 0
Aj 5 , C j 2
A 3，
a
m2 0 1 4i m 0
2 j
A03 .
m2 j 1 4i m j
Aj 6 ,
(40)
CR IP T
C 1
a
The unknowns A 1 ~ A 3 , Aj1 ~ Aj 6 and A01 ~ A03 could be solved by the boundary conditions. Then,
AN US
q , T and p can be expressed as
q q x, h, t a
a
2
m2 1 4im
A 3ebt i mx ,
M
aN2 m 2 N 1 T TN ( x, h, t ) T ( x, h, t ) 4i m N , 2 2 a m 1 A ebt i mx AN 5e aN h AN 6 e aN h 3 4i m
ED
(41)
PT
1 p y ( x, h, t ) 2i mA 1 A 2 mA 3 ebt i mx . 2
CE
3.3 Boundary conditions 3.3.1 The FGM coating
AC
The displacement field, temperatures, heat flux and stress field are continued at the interface of the
sub-layers, e.g.,
uxj x, h j , t ux, j +1 x, h j , t ，u yj x, h j , t u y , j 1 x, h j , t , Tj x, h j , t T j 1 x, h j , t ，q j x, h j , t q j 1 x, h j , t ,
yj x, h j , t y , j 1 x, h j , t ， xyj x, h j , t xy , j 1 x, h j , t . Then, Eq. (42) can be written in the matrix form 11
(42)
ACCEPTED MANUSCRIPT j h j j j 1 h j j 1 ,
where j Aj1 , Aj 2 , Aj 3 , Aj 4 , Aj 5 , Aj 6
T
(43)
and
j y 1 , 2 , 3 , 4 , 5 , 6 , T
1 = emy，yemy，e my，ye my，e
a j y
，e
aj y
,
j my my j my a j a y a j a y 2 = emy， e , y e ，e ， ye ， e ， j
m
m
m
m
CR IP T
j
a 2j m2 j 1 a j y a 2j m2 j 1 a j y = 0, 0, 0, 0， e ， e , j j 3
a 2j m2 j 1 a j y a 2j m2 j 1 a j y = 0, 0, 0, 0, a j j e , a j j e , j j
=2m e
AN US
4
5 = 2m j emy , j 2my j 1 emy , 2m j e my , j 2my j 1 emy , 2m j e
a j y
, 2m j e
, j 2my j 1 emy , 2m j e my , j 2my j 1 e my , 2 j a j e
a j y
, 2 j a j e
6
j
my
aj y
is the transfer matrix. Finally, we have
M
Thus, j 1 j j , where j j11 h j j h j
, .
aj y
(44)
ED
N N 1N 2 ...11 .
3.3.2 At y 0
PT
In order to maintain the integrity of the FGM coating and bottom homogeneous half-plane, we
CE
assume the boundary conditions at y 0 to be continuous
AC
ux 0 x,0, t ux1 x,0, t ，u y 0 x,0, t u y1 x,0, t , T0 x,0, t T1 x,0, t ，q0 x,0, t q1 x,0, t ,
(45)
y 0 x,0, t y1 x,0, t ， xy 0 x,0, t xy1 x,0, t .
Eq. (45) can be expressed in the matrix form 00 =1 h0 1 , then
1 =1-1 h0 00 , where 0 A01 , A02 , A03 and T
12
(46)
ACCEPTED MANUSCRIPT 0 01 ,02 ,03 , 01 1, 1， 0， 0， 2m0 , 2m0 , T
0， 0 , 0， 0, 0 0 1 , 0 1 0 , m T
2 0
2 2 a02 m2 0 1 a0 a0 m 0 1 1， , , a00 , 2m0 , 2a0 0 . m 0 0 T
CR IP T
3 0
3.3.3 At y h
At the sliding surface ( y h ), the displacements and stresses are continuous [28, 45], i.e.,
AN US
u y x, h, t u yN x, h, t ,
(47)
y x, h, t yN x, h, t ,
(48)
xy x, h, t 0 ,
(49)
M
(50)
where A 1 , A 2 , A 3
= N N
(51)
and
PT
T
ED
Similarity, the matrix form of Eqs. (47)-(50) and (4-5) is
1 ,2 ,3 ,
1 N 1 N * * * 1, 2m , 2m, 0, 2mV * 1 , m Q 2 R V 1 , N 1 N 1 T
CE
AC
1
1 N * 1 1 N * * * , 1 ,1 , 0, 1 1 V , 1 Q 2 R V , 2 N 1 m N 1 T
2
2 2 a 1 N a a m * , 2m , 2a , 0, 2mV 1 , 2 m 1 m N 3
2 2 1 N a m a * * * * m Q 2 R V 1 1 R , 1 m m N T
N N1 , N2 , N3 , N4 , N5 , N6 , T
13
ACCEPTED MANUSCRIPT
N1 emh , h
N
N mh mh e ,e , h m m
mh aN aN h aN aN h e , e , e , m m
N2 2mN emh , N 1 N 2mh emh , 2mN e mh , N 1 N 2mh e mh , 2mN e a h , 2mN ea N
Nh
,
N3 = 0,0,0,0,0,0 , N4 2memh , 2mh N 1 emh , 2me mh , 2mh N 1 e mh , 2aN e a h , 2aN ea N
Nh
,
aN N N 1 aN2 m2 a h aN N N 1 aN2 m2 a h 0, 0, 0, 0, e N , e N , 2 2 m N 1 m N 1
CR IP T
5 N
N 1 aN2 m2 a h N 1 aN2 m2 a h 0, 0, 0, 0, e N , e N , m 1 m 1 N N 6 N
with
TN 0 T 0 1 1 1 1 N , V * =2 fMV , . N R0 mk M 2
AN US
R* mR0 k , Q* 4 RM
Note that R*，Q* and V * are respectively the dimensionless thermal contact resistance, interface
M
temperature drop or heat flux and sliding speed at the interface [28].
ED
From Eqs. (44), (46) and (51), we obtain
0 .
(52)
PT
where N N 1 N 2 ...111 h0 0 . Eq. (52) can be expressed in the dimensionless form as
* *0 ,
(53)
CE
by utilizing the following dimensionless parameters
AC
a0*
a0 a b b 1+ 2 1 z , a* 1 2 1 z , m m 0 0 m m
a*j
aj m
1
b m j 2
1
b z , z 2 , H mh , m j
(54)
(55)
where * and * are the dimensionless forms of and . It is easy to know that * and * are 6 3 matrix, so Eq. (52) can be treated as six homogeneous linearity equations with six unknowns
A01 , A02 , A03 , A 1 , A 2 and A 3 . If and only if the coefficient matrix equals to zero, there are nontrival solutions for these six equations. That is to say, 14
ACCEPTED MANUSCRIPT det * , * 0 .
(56)
Eq. (56) is the characteristic equation of the exponential growth rate z
z b / m . 2
Eq. (56)
includes the interaction of the friction coefficient and the thermal contact resistance, which makes it possible for us to study the coupled effects of them on the stability boundaries meanwhile.
CR IP T
4. Stability criterion It will cause the system to be unstable if the assumed form perturbation has a dimensionless
exponential growth rate z z b / m2
with real and positive or complex with positive real part [20].
To study the stability boundary, we consider two different kinds of the exponential growth z.
AN US
Firstly, z is real and z 0 . The characteristic equation (56) turns to the linear relation between the critical sliding speed V * and critical heat flux Q* . Secondly, z is complex with Re[z] 0 . There are three unknowns in the characteristic equation (56), i.e., the critical sliding speed V * , critical heat flux
Q* and real w. Separating the real and imaginary parts of Eq. (56), we have
(57)
Im det * , * 0 ,
(58)
ED
M
Re det * , * =0 ，
PT
from which we can have the relationship between V * and Q* via the parameter w.
CE
5. Results and discussion
In the sliding systems like brakes and clutches, the frictional heat and thermal contact resistance
AC
always exist at the same time. They will lead to the non-uniform thermal distortions and thermal distribution, which ultimately induce one of the commonest and most dangerous appearances, i.e., the hot spots. It is not accurate enough only considering one of them. In this section, we will focus on the thermoelastic contact instability between a homogeneous half-plane and an FGM coated half-plane with arbitrarily varying properties, taking both the frictional heat and thermal contact resistance into account. In this paper, we mainly study the effects of the thermal diffusivity coefficient , the gradient index, the thickness of the FGM coating H, the thermal contact resistance R* , the friction coefficient f and the heat partition coefficient on the stability boundaries. The FGM coating is made by a mixture of the ceramic 15
ACCEPTED MANUSCRIPT and nodular cast iron. The material of the FGM coating at y 0 is nodular cast iron. The material of the bottom homogeneous half-plane is also taken as the nodular cast iron. The material of the upper homogeneous half-plane is made by the friction material [40]. The properties of the metal, ceramic and friction materials are listed in Table 1 [20, 40]. Since the homogeneous multi-layered model can simulate FGMs with arbitrarily varying properties, we consider FGMs with the power-law form, exponential function, sinusoidal function and cosine function,
the power-law function through the thickness direction, n
1 y
y 0 N 0 , h n
AN US
2 y k y k0 k N k0 , h
CR IP T
respectively. We name the power-law form as Case 1 in which the properties of the FGM coating follow
(59a)
(59b)
n
3 y y 0 N 0 , h
y
n4
M
y 0 N 0 , h
ED
PT CE AC
(59d)
n5
y y 0 N 0 , h
the exponential form
(59c)
(59e)
y 0e y , 1 ln N / 0 / h ,
(60a)
k y k0e2 y , 2 ln kN / k0 / h ,
(60b)
y 0e y , 3 ln N / 0 / h ,
(60c)
y 0e y , 4 ln N / 0 / h ,
(60d)
y 0e y , 5 ln N / 0 / h ,
(60e)
1
3
4
5
as Case 2; the sinusoidal form
y , 2h
y 0 N 0 sin
16
(61a)
ACCEPTED MANUSCRIPT y k y k0 k N k0 sin , 2h
(61b)
y , 2h
(61c)
y , 2h
(61d)
y , 2h
(61e)
y 0 N 0 sin
y 0 N 0 sin as Case 3 and the cosine form
0 N 2
0 N 2
2
y
0 N 2
0 N 2
0 N
M
0 N
ED
y
2
y cos , h
k0 k N k 0 k N y cos , 2 2 h
k y
y
0 N
AN US
y
CR IP T
y 0 N 0 sin
2
0 N 2
(62a)
(62b)
y cos , h
(62c)
y cos , h
(62d)
y cos , h
(62e)
PT
as Case 4, where N , kN , N , N and N are the shear modulus, thermal conductivity coefficient,
CE
thermal expansion coefficient, thermal diffusivity coefficient and Poisson’s radio of the upper surface of
AC
the FGM coating, respectively.
5.1 Convergence and comparison studies Before analyzing the TEI of FGMs, we need to examine the convergence of the homogeneous
multi-layered model. Table 2 shows the effect of the total number of sub-layers (N) on the critical sliding speed V * with H =1.0, R* = 1.0, f = 0.3, Q* 30 and 0.5 . Note that both n 0.1 and stand for the greatly varying material properties of FGM.
n 10
It is observed that the critical sliding speeds
become closer to each other with an increasing N. Obviously, the accurate results can be obtained if the 17
ACCEPTED MANUSCRIPT number of sub-layers N = 8 or 10 for n 1.0 , and N = 12 or 14 for n 0.1 and 10 . Therefore, we choose 12 sub-layers for all examples in the next analysis. It should be pointed out that the extreme gradient slop of FGMs, for example n 0.01 and
n 100 , are beyond the consideration of the
present analysis. Actually, it is hard to fabricate FGMs with the extreme gradient slop. So, the gradient index is limited at the range 0.1 n 10 in our discussion. We choose the stainless and aluminum alloy, respectively, as the material of the upper homogeneous
CR IP T
and bottom homogeneous half-plane. The present problem can be reduced to the static TEI problem [20] if we set the sliding speed and the thickness of the FGM coating as zero. Fig. 2 presents the stability boundaries ( Q* vs R*) for a homogeneous stainless steel half-plane and a homogeneous aluminum alloy half-plane with V * 0.0 . Zhang and Barber’s results are also plotted in Fig. 2 to validate the present
AN US
analysis. Obviously, the present results are in good agreement with Zhang and Barber’s results [20]. If we neglect the sliding speed and bottom homogeneous half-plane, the present coupled TEI problem can be reduced to the static TEI problem between and FGM layer and a homogeneous half-plane reported by Mao et al. [2]. Fig. 3 compares the stability boundaries of the static TEI problem by using the
M
homogeneous multi-layered model and exponential model with H=1.0 and n =0.2. The FGM coating is made by a mixture of the ceramic and nodular cast iron. Again, it can be seen that multi-layered model’s
ED
results agree well with exponential model’s results [2].
PT
5.2 The results for the power-law case
For the sake of convenience, we assume that the gradient indices have the same value n in Eq. (59),
CE
i.e., n1 n2 n3 n4 n5 n . Fig. 4 depicts the effect of the gradient index n on the stability
AC
boundaries with H =1.0, R* = 1.0, f = 0.3 and 0.5 . The stability boundary depends on the real root first and then on the complex root for n = 0.5, 1.0 and 2.0. In the present analysis, we call a special point as the turning point where the instability changes from the real root to the complex root. We can see that the critical sliding speed at the turning point decreases with the decrease of the gradient index n, for example V * 370 for n = 2.0, V * 310 for n = 1.0 and V * 240 for n = 0.5. For the real root instability, the relationship between V * and Q* is linear. However, it is no longer linear for the complex root instability. The same phenomenon is also found by Afferante et al. [28]. For the positive heat 18
ACCEPTED MANUSCRIPT flux, the critical sliding speed decreases with the increase of Q* . But the trend is opposite for the negative heat flux for which the critical sliding speed increases with the increase value of Q* . In other words, the frictional heat can make the system stable in certain situations when considering the pressure-dependent thermal contact resistance [29]. With the increase of the gradient index, the stability area becomes larger. That is because the bigger the gradient index is, the upper sub-layers of the FGM
CR IP T
coating are the ceramic-rich layers, which is more easier to keep stability than the metal-rich layers. Fig. 5 examines the effect of the thickness H on the stability boundaries with n =1.0, R* = 1.0, f = 0.3 and 0.5 . The stability boundaries are all relied on the real roots first and then complex roots for the given thickness H = 1.0, 2.0 or 3.0. Similar to Fig. 3, the relationship between V * and Q* is no longer
AN US
linear for the complex root instability, and the positive heat flux unstabilizes the system, while the negative heat flux stabilizes the system. It is found that the critical speed at the turning point decreases with the increase of H. The dimensionless critical speed is 310 for H = 1.0, 190 for H = 2.0 and 140 for H = 3.0. Note that this conclusion depends on the combinations of coating and substrate materials and the
M
ratios of the material parameters at the bottom and upper surfaces of the coating. If we take other combinations of coating and substrate materials, the conclusion may be opposite from the current one.
H = 1.0, f = 0.3 and
ED
Fig. 6 explores the effect of the thermal contact resistance R* on the stability boundaries with n =1.0,
0.5 . There is still a critical sliding speed V * at the turning point for each curve,
PT
below which the stability boundaries are decided on the real roots, and above which the boundaries depend on the complex roots. Interestingly, the critical sliding speed at the turning points has the same
CE
value 310 for three different R* 0.5, 0.8, 1.0. That is to say, the thermal contact resistance has no
AC
effect on the turning point. However, it has a great effect on the stability area, which decreases with the increase of the thermal contact resistance. The bigger the thermal contact resistance is, the more the frictional heat accumulated at the interface. So the system is much more possible to be unstable. Fig. 7 presents the effect of the friction coefficient f on the stability boundaries with n =1.0, H = 1.0, R*
= 1.0 and
=0.5 . It can be seen that the critical speed V * at the turning point increases with the
decrease of the friction coefficient f, e.g. V * 310 for f = 0.30, V * 370 for f = 0.25 and V * 465 for f = 0.20. The increase of the friction coefficient f leads to the decrease of the stability area. In the 19
ACCEPTED MANUSCRIPT sliding system, the frictional heat is one of the most important factors for the TEI. The bigger the friction coefficient is, the more the heat generated by the friction is. Thus, the system is much more possible to be unstable with the bigger friction coefficient f. Fig. 8 shows the effect of the heat partition coefficient on the stability boundaries with n =1.0, H = 1.0, R* = 1.0 and f = 0.30. It is extremely difficult to confirm the value because it is closely relevant
CR IP T
to the mechanism of the interface. Unless there are available data obtained from the experiments, researchers always select the value within the range of the possible values 0.0~1.0 [28]. Therefore, it is chosen as 0.2, 0.5 and 0.8 in the discussion. It is observed that the turning point is insensitive to the heat partition coefficient. However, the stability area is increasing by the decreasing heat partition coefficient. According to the definition of the heat partition coefficient in section 2, the smaller heat partition
AN US
coefficient indicates that small amount of the frictional heat flows into the upper homogeneous half-plane, but large amount of the frictional heat flows into the FGM coated half-plane.
5.3 Compare the results for different gradient types
M
It is known that the homogeneous multi-layered model has the merit to simulate the arbitrarily
ED
varying material properties of FGMs. By using this model, it is possible to check the effect of different gradient types of FGMs on the stability behavior of system. Indeed, we can select a lot of types of distribution in the analysis. For simplicity, we only select four gradient types of FGMs, i.e, the power-law
PT
function, exponential function, sinusoidal function and cosine function.
CE
Fig. 9 illustrates the stability boundaries for the four different cases with H = 1.0, R* = 1.0, f = 0.3 and
0.5 . For Case 1, the gradient index is takes as n 1.0 . Among the four cases, it is found that Case 1
AC
has the largest stability area and the critical sliding speed at the turning point. The difference of results obtained from different gradient types is great for the complex root instability, while it is relatively small for the real root instability. These results indicate that an appropriate gradient type of the FGM coating can adjust the TEI of systems.
5.4 Discuss the effect of the thermal diffusivity coefficient In previous works concerning the thermoelastic crack problem of FGMs [47, 48], thermoelastic contact of FGMs [36] and TEI of FGMs [40-45], the thermal diffusivity coefficient is always treated as a 20
ACCEPTED MANUSCRIPT constant for the mathematical convenience. In fact, the thermal diffusivity coefficient is inhomogeneous along the thickness direction. Therefore, this treatment of the thermal diffusivity coefficient may lead to the error in the TEI analysis of FGMs. Firstly, we discuss the effect of the thermal diffusivity coefficient on the TEI of FGMs considering the coupled effect of the frictional heat and thermal contact resistance. The properties of the FGM coating follow the power-law function (i.e., case 1). Tables 3 presents the effect of the different form of the
CR IP T
thermal diffusivity coefficient (the graded and the constant ) on the critical sliding speed V * with H =1.0, R* = 1.0, f = 0.3 and 0.5 . Note that 0 and N are the thermal diffusivity coefficient of FGM coating at the bottom surface and upper surface, respectively. “Graded with n =1.0” denotes that the thermal diffusivity coefficient is inhomogeneous along the thickness direction. “ 0 ”,
AN US
“ (0 +N ) 2 ” and “ (0 +N ) 2 ” denote the ways to treat the thermal diffusivity coefficient as constant. The homogeneous multi-layered model is used to solved the TEI of FGMs with the graded or the constant . The values in the bracket denote the relative difference percentages between the solutions of the constant and graded . It is defined as the absolute value of (graded solution -
M
constant solution) / (graded solution). It can be observed that there is no difference between the
ED
graded thermal diffusivity coefficient and the constant for the real root instability. However, for the complex root instability, the different forms of the thermal diffusivity coefficient have a great effect on the
PT
critical sliding speed.
Secondly, we want to examine the error in our previous work about the TEI of FGMs [2, 41-43] with
CE
the thermoelastic parameters varying in the exponential form. In those works, the thermal diffusivity coefficient is treated as a constant by assuming that the gradient indices of the thermal conductivity,
AC
density and specific heat satisfy a certain relation [2, 40-45]. Here, we only consider the effect of the different form of on the critical siding speed V * for the complex root instability since it indeed has no effect on the real root instability behavior. Table 4 tabulates the effect of the thermal diffusivity coefficient on the TEI of FGMs with their properties varying exponentially (i.e., case 2). For the graded , we assume that the gradient indexes have the same value
in Eq. (60), i.e.,
1 2 3 4 5 . For the constant , we have 1 2 3 5 and j 0 or
21
ACCEPTED MANUSCRIPT j N . The bottom surface of the FGM coating is chosen as the nodular cast iron. The gradient index
is respectively selected as 0.1, 0.2, 0.5 and 0.6. It is found that the error is about 0.03%~3.74% between the constant and graded when the gradient index 0.6 . This example indicates that the assumption of constant is reasonable in our previous works [2, 42, 43, 45] when the gradient index
CR IP T
of the exponential function is small.
6. Conclusions
This paper presents the TEI of FGMs considering the coupled effect of the thermal contact resistance and frictional heat in the plane strain state through the perturbation method and the transfer matrix
AN US
method. A homogeneous half-plane slides against an FGM coated half-plane at the out of plane direction. The properties of the FGM coating vary arbitrarily. The homogeneous multi-layered model is used to deal with the arbitrarily varying properties of the FGM coating. The effects of the gradient index, different gradient types and the thermal diffusivity coefficient of the FGM coating on the stability boundaries are discussed. It is found that：
M
(1) For the power law case, the stability boundaries grow with the thermal contact resistance, friction
ED
coefficient, heat partition coefficient and the increasing gradient index of the FGM coating. (2) An appropriate gradient type and thickness of the FGM coating can adjust the TEI of systems. (3) For the power law case, the different form of the thermal diffusivity coefficient of FGM coating has
PT
no effect on the real root instability, but the effect is great for the complex stability behavior. Interestingly,
CE
the results obtained from the average value of the thermal diffusivity coefficient are very close to those obtained from the graded thermal diffusivity coefficient.
AC
(4) For the exponential case, the assumption of the constant thermal diffusivity coefficient is reasonable for the TEI of FGMs when the gradient index is small.
Acknowledgements The work described in this paper is supported by the Fundamental Research Funds for the Central Universities under Grant number 2016YJS106.
References 22
ACCEPTED MANUSCRIPT [1] Barber JR. Thermoelasticity and contact. Journal of Thermal Stresses 1999;22(4-5): 513-25. [2] Mao JJ, Ke LL, Wang YS. Thermoelastic contact instability of a functionally graded layer and a homogeneous half-plane. International Journal of Solids and Structures 2014;51(23): 3962-72. [3] Barber JR. Thermoelastic instabilities in the sliding of comforming solids. In: Proceedings of the Royal Society of London A: Mathematical, Physical and Engineering Sciences. The Royal Society: 1969. p.381-94.
CR IP T
[4] Dow TA, Burton RA. Thermoelastic instability of sliding contact in the absence of wear. Wear 1972;19(3):315-28.
[5] Burton RA, Nerlikar V, Kilaparti SR. Thermoelastic instability in a seal-like configuration. Wear 1973;24(2):177-88.
AN US
[6] Lee K, Barber JR. Frictionally excited thermoelastic instability in automotive disk brakes. Journal of Tribology 1993;115(4):607-14.
[7] Lee K. Frictionally excited thermoelastic instability in automotive drum brakes. Journal of Tribology 2000;122(4):849-55.
[8] Decuzzi P, Ciaverella M, Monno G. Frictionally excited thermoelastic instability in multi-disk clutches
M
and brakes. Journal of Tribology 2001;123(4):865-71.
ED
[9] Du S, Zagrodzki P, Barber JR, Hulbert GM. Finite element analysis of frictionally excited thermoelastic instability. Journal of Thermal Stresses 1997;20(2):185-201. [10] Yi YB, Du S, Barber JR, Fash JW. Effect of geometry on thermoelastic instability in disk brakes and
PT
clutches. Journal of Tribology 1999;121(4):661-6.
CE
[11] Yi YB, Barber JR, Zagrodzki P. Eigenvalue solution of thermoelastic instability problems using Fourier reduction. In: Proceedings of the Royal Society of London A: Mathematical, Physical and
AC
Engineering Sciences. The Royal Society: 2000:2799-821. [12] Audebert N, Barber JR, Zagrodzki P. Buckling of automatic transmission clutch plates due to thermoelastic/plastic residual stresses. Journal of Thermal Stresses 1998;21(3-4):309-26.
[13] Panier S, Dufrenoy P, Bremond P. Infrared characterization of thermal gradients on disc brakes. In: AeroSense 2003:295-302. [14] Afferrante L, Ciavarella M, Barber JR. Sliding thermoelastodynamic instability. Proceedings of the Royal Society of London A: Mathematical, Physical and Engineering Sciences. 2006;462(2071): 2161-2176. 23
ACCEPTED MANUSCRIPT [15] Afferrante L, Ciavarella M. Thermo-elastic dynamic instability (TEDI) in frictional sliding of two elastic half-spaces. Journal of the Mechanics and Physics of Solids. 2007; 55(4):744-764. [16] Afferrante L, Ciavarella M. Thermoelastic dynamic instability (TEDI) in frictional sliding of a half-space against a rigid non-conducting wall. 2007; 74(5): 875-884. [17] Afferrante L, Ciavarella M. Thermo-elastic dynamic instability (TEDI)—a review of recent results. Journal of Engineering Mathematics. 2008; 61:285-300.
CR IP T
[18] Zelentsov VB, Mitrin BI, Aizikovich SM. Dynamic and quasi-static instability of sliding thermoelastic frictional contact. Journal of Friction and Wear. 2016;37(3):213-220.
[19] Barber JR. Stability of thermoelastic contact. In: Proceeding Institution of Mechanical Engineers. International Congress on Tribology 1987:981-86.
AN US
[20] Zhang R, Barber JR. Effect of material properties on the stability of static thermoelastic contact. Journal of Applied Mechanics 1990;57(2):365-9.
[21] Yeo T, Barber JR. Stability of thermoelastic contact of a layer and a half-plane. Journal of Thermal Stresses 1991;14(4):371-88.
Thermal Stresses 1997;20(2):169-84.
M
[22] Li C, Barber JR. Stability of thermoelastic contact of two layers of dissimilar materials. Journal of
ED
[23] Schade DT, Oditt K, Karr DG. Thermoelastic stability of two bonded half planes. Journal of Engineering Mechanics 2000;126(9):981-5. [24] Schade DT, Karr DG. Thermoelastic stability of layer bonded to half plane. Journal of Engineering
PT
Mechanics 2002;128(12):1285-94.
CE
[25] Johansson L. Model and numerical algorithm for sliding contact between two elastic half-planes with frictional heat generation and wear. Wear 1993;160(1):77-93.
AC
[26] Ciavarella M, Johansson L, Afferrante L, Klarbring A, Barber JR. Interaction of thermal contact resistance and frictional heating in thermoelastic instability. International Journal of Solids and Structures 2003;40(21):5583-97.
[27] Afferrante L, Ciavarella M. Frictionally excited thermoelastic instability in the presence of contact resistance. International Journal of Solids and Structures 2004;39(4): 351-7. [28] Afferrante L, Ciavarella M. Instability of thermoelastic contact for two half-planes sliding out-of-plane with contact resistance and frictional heating. Journal of the Mechanics and Physics of Solids 2004;52(7):1527-47. 24
ACCEPTED MANUSCRIPT [29] Ciavarella M, Barber JR. Stability of thermoelastic contact for a rectangular elastic block sliding against a rigid wall. European Journal of Mechanics-A/Solids 2005;24(3):371-6. [30] Suresh S, Mortensen A. Functionally graded metals and metal-ceramic composites: Part 2 Thermomechanical behaviour. International Materials Reviews 1997;42(3):85-116. [31] Krenev LI, Aizikovich SM, Tokovyy YV, Wang YC. Axisymmetric problem on the indentation of a hot circular punch into an arbitrarily nonhomogeneous half-space. International Journal of Solids and
CR IP T
Structures 2015;59:18-28. [32] Chen P, Chen S, Peng Z. Thermo-contact mechanics of a rigid cylindrical punch sliding on a finite graded layer. Acta Mechanica 2012;223(12):2647-65.
[33] Chen P, Chen S. Partial slip contact between a rigid punch with an arbitrary tip-shape and an elastic
AN US
graded solid with a finite thickness. Mechanics of Materials 2013;59:24-35.
[34] Suresh S. Graded materials for resistance to contact deformation and damage. Science 2001;292(5526):2447-51.
[35] Eshraghi I, Dag S, Soltani N. Bending and free vibrations of functionally graded annular and circular micro-plates under thermal loading. Composite Structures 2016 Mar 31;137:196-207.
involving
2011;48(18):2536-48.
frictional
heating.
International
Journal
of
Solids
and
Structures
ED
materials
M
[36] Liu J, Ke LL, Wang YS. Two-dimensional thermoelastic contact problem of functionally graded
[37] Kulchytsky-Zhyhailo R, Bajkowski AS. Axisymmetrical problem of thermoelasticity for halfspace
PT
with gradient coating. International Journal of Mechanical Sciences 2016 Feb 29;106:62-71.
CE
[38] Chen PJ, Chen SH, Yao Y. Nonslipping contact between a mismatch film and a finite-thickness graded substrate. Journal of Applied Mechanics 2016;83(2):021007.
AC
[39] Liu J, Ke LL, Wang YS, Yang J, Alam F. Thermoelastic frictional contact of functionally graded materials with arbitrarily varying properties. International Journal of Mechanical Sciences 2012;63(1):86-98.
[40] Lee SW, Jang YH. Effect of functionally graded material on frictionally excited thermoelastic instability. Wear 2009;266(1):139-46. [41] Lee SW, Jang YH. Frictionally excited thermoelastic instability in a thin layer of functionally graded material sliding between two half-planes. Wear 2009;267(9):1715-22. [42] Mao JJ, Ke LL, Wang YS. Thermoelastic instability of functionally graded materials in frictionless 25
ACCEPTED MANUSCRIPT contact. Acta Mechanica 2015; 226: 2295-311. [43]Mao JJ, Ke LL, Wang YS. Thermoelastic instability of a functionally graded layer interacting with a homogeneous layer. International Journal of Mechanical Sciences 2015; 99:218-27. [44]Liu J, Wang YS, Ke LL, Yang J, Alam F. Dynamic instability of an elastic solid sliding against a functionally graded material coated half-plane. International Journal of Mechanical Sciences 2014;89:323-31.
materials
sliding
out-of-plane
with
contact
2016;83(2):021010.
resistance.
CR IP T
[45] Mao JJ, Ke LL, Wang YS, Liu J. Frictionally excited thermoelastic instability of functionally graded Journal
of
Applied
Mechanics
[46] Li C. Thermoelastic contact stability analysis, Ph.D. thesis. University of Michigan, Michigan: 1998.
AN US
[47] Jin ZH, Noda N. Transient thermal stress intensity factors for a crack in semi-infinite plate of a functionally gradient material. International Journal of Solids and Structures 1994; 31:203-18. [48] Shodja HM, Ghahremaninejad A. An FGM coated elastic solid under thermomechanical loading: a
AC
CE
PT
ED
M
two dimensional linear elastic approach. Surface and Coatings Technology 2006; 200(12):4050-64.
26
ACCEPTED MANUSCRIPT Table captions
Table 1 Thermoelastic properties of selected materials Table 2 The effect of the total number of multi-layers (N) on the critical sliding speed V* for different gradient types Table 3 The effect of the thermal diffusivity coefficient on the critical sliding speed V*: Power-law
CR IP T
function (Case 1) Table 4 The effect of the thermal diffusivity coefficient on the critical sliding speed V* with Q* = -180:
AC
CE
PT
ED
M
AN US
exponential function (Case 2)
27
ACCEPTED MANUSCRIPT Figure captions
Fig. 1 Schematic map of the thermoelastic contact problem for a homogeneous half-plane sliding out of an FGM coated half-plane (a) and the homogeneous multi-layered model (b). Fig. 2 The stability boundaries for a homogeneous stainless half-plane and a homogeneous aluminum alloy half-plane with V* = 0.0 with R* = 1.0, f = 0.3 and 0.5 .
CR IP T
Fig. 3 The stability boundaries of the static TEI problem with H = 1.0 and n 0.2 .
Fig. 4 The effect of gradient index n on the stability boundaries with H =1.0, R* = 1.0, f = 0.3 and
0.5 .
Fig. 5 The effect of the thickness H on the stability boundaries with n =1.0, R* = 1.0, f = 0.3 and 0.5 .
AN US
Fig. 6 The effect of thermal contact resistance R* on the stability boundaries with n =1.0, H = 1.0, f = 0.3 and 0.5 .
Fig. 7 The effect of the friction coefficient f on the stability boundaries with n = 1.0, H = 1.0, R* = 1.0
M
and 0.5 .
Fig. 8 The effect of heat partition coefficient on the stability boundaries with n =1.0, H = 1.0, R* = 1.0
ED
and f 0.3 .
AC
CE
PT
Fig. 9 The stability boundaries for the four different cases with H = 1.0, R* = 1.0, f = 0.3 and =0.5 .
28
ACCEPTED MANUSCRIPT
Table 1 Thermoelastic properties of selected materials Friction
Nodular
Ceramic
Aluminium
Stainless
material
cast iron
alloy
steel
0.3
168
151
72
14.0
13.7
12.0
22.0
k (W/moC)
0.241
48.9
3.0
( mm2/s)
0.177
16.05
1.15
0.12
0.31
SiC
Properties
CR IP T
6
0.30
AC
CE
PT
ED
M
o C1 10
200
400
14.0
4.4
173.0
21.0
110.0
67.16
5.93
35.48
0.30
0.16
AN US
E (GPa)
29
sintered
0.32
ACCEPTED MANUSCRIPT
Table 2 The effect of the total number of multi-layers (N) on the critical sliding speed V* for different gradient types
6
8
10
12
14
16
n = 0.1
93.07
94.44
95.02
95.35
95.55
95.70
95.80
n = 1.0
100.06
102.86
103.93
104.47
104.80
105.02
105.17
n = 10.0
101.81
105.29
106.65
108.07
108.27
AN US
4
CR IP T
Total number of multi-layers (N) n
AC
CE
PT
ED
M
107.36
30
107.79
ACCEPTED MANUSCRIPT
Table 3 The effect of the thermal diffusivity coefficient on the critical sliding speed V*: Power-law function (Case 1)
2
-30
6.40
104.80
Graded with
0
6.40
N
6.40
6.40 Real
347.11
398.96
320.57
333.25
(7.65%)
(16.47%)
343.44
390.17
(1.06%)
(2.20%)
379.29
491.42
(9.27%)
(23.18%)
Complex
Complex
104.80
104.80 Real
AC
CE
PT
ED
Instability Types
-160
104.80
M
(0 +N ) 2
-120
AN US
n =1.0
CR IP T
Q*
31
ACCEPTED MANUSCRIPT
Table 4 The effect of the thermal diffusivity coefficient on the critical sliding speed V* with Q* = -180:
0.2
0.5
0.6
Graded
592.72
585.30
540.55
518.40
0
592.87
586.47
550.12
533.27
(0.03%)
(0.20%)
(1.77%)
(2.87%)
AN US
0.1
592.54
N
CR IP T
exponential function (Case 2)
(0.03%) Instability Types
583.98
527.92
499.00
(0.23%)
(2.33%)
(3.74%)
AC
CE
PT
ED
M
Complex
32
ACCEPTED MANUSCRIPT
AN US
CR IP T
Interface
AC
CE
PT
ED
M
(a)
(b)
Fig. 1 Schematic map of the thermoelastic contact problem for a homogeneous half-plane sliding out of an FGM coated half-plane (a) and the homogeneous multi-layered model (b).
33
ACCEPTED MANUSCRIPT
30
0
CR IP T
Q
*
-30
Present results: Real root Complex root
-60
-90
-120 0.0
0.2
0.4
AN US
Zhang and Barber's results: Real root Complex root
*
0.6
0.8
1.0
1/(1+R )
M
Fig. 2 The stability boundaries for a homogeneous stainless half-plane and a homogeneous aluminum
AC
CE
PT
ED
alloy half-plane with V* = 0.0 with R* = 1.0, f = 0.3 and 0.5 .
34
ACCEPTED MANUSCRIPT
120
Exponential model [2] Multi-layered model
CR IP T
105
Q
*
90
75
AN US
60
45 0.00
0.15
0.30
0.45
0.60
*
M
1/(1+R )
AC
CE
PT
ED
Fig. 3 The stability boundaries of the static TEI problem with H = 1.0 and n 0.2 .
35
ACCEPTED MANUSCRIPT
20 0
CR IP T
-20 -40
*
Q
-80 -100 -120 -140
Instability
Stability
-160 0
50
100
150
AN US
Real roots: n = 0.5 n = 1.0 n = 5.0 Complex roots n = 0.5 n = 1.0 n = 5.0
-60
200
250
M
V
300
350
400
450
*
0.5 .
AC
CE
PT
ED
Fig. 4 The effect of the gradient index n on the stability boundaries with H =1.0, R* = 1.0, f = 0.3 and
36
ACCEPTED MANUSCRIPT
20 0
Real roots: H = 1.0 H = 2.0 H = 3.0
CR IP T
-20
Instability
-40
-80
Complex roots: H = 1.0 H = 2.0 H = 3.0
-100 -120 -140
Stability
-160 0
50
100
AN US
Q
*
-60
150
200
250
300
350
400
*
M
V
AC
CE
PT
ED
Fig. 5 The effect of the thickness H on the stability boundaries with n =1.0, R* = 1.0, f = 0.3 and 0.5 .
37
ACCEPTED MANUSCRIPT
20 0 -20 -40
CR IP T
Instablity Real roots: * R = 0.5 * R = 0.8 * R = 1.0
Q
*
-60 -80 -100
Complex roots: * R = 0.5 * R = 0.8 * R = 1.0
-140
Stablity
AN US
-120
-160 0
50
100
150
200
250
300
350
400
*
M
V
Fig. 6 The effect of the thermal contact resistance R* on the stability boundaries with n =1.0, H = 1.0, f =
AC
CE
PT
ED
0.3 and 0.5 .
38
ACCEPTED MANUSCRIPT
20 0 -20
Complex roots: f = 0.30 f = 0.25 f = 0.20
CR IP T
Instability
-40
Real roots: f = 0.20 f = 0.25 f = 0.30
-80 -100 -120
AN US
Q
*
-60
Stability
-140 -160 0
50
100
150
200
250
300
350
400
450
500
*
M
V
0.5 .
AC
CE
PT
ED
Fig. 7 The effect of the friction coefficient f on the stability boundaries with n = 1.0, H = 1.0, R* = 1.0 and
39
ACCEPTED MANUSCRIPT
30 0
Real roots:
*
-60
Q
CR IP T
Instability
-30
-90 -120
Complex roots:
-180
Stability
AN US
-150
-210 0
50
100
150
200
250
300
350
400
*
M
V
ED
Fig. 8 The effect of heat partition coefficient on the stability boundaries with n =1.0, H = 1.0,
AC
CE
PT
R* 1.0 and f = 0.3.
40
ACCEPTED MANUSCRIPT
20 0
Real roots: Case 1 (n = 1.0) Case 2 Case 3 Case 4
CR IP T
Instability
-20 -40
Stability
-80
Complex roots: Case 1 (n = 1.0) Case 2 Case 3 Case 4
-100 -120 -140 -160 0
50
100
150
AN US
Q
*
-60
200
250
300
350
400
*
M
V
AC
CE
PT
ED
Fig. 9 The stability boundaries for the four different cases with H = 1.0, R* = 1.0, f = 0.3 and =0.5 .
41