See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/267456625

Surface tension-induced stress concentration around a nanosized hole of

arbitrary shape in an elastic half-plane

Article  in  Meccanica · December 2014

DOI: 10.1007/s11012-014-0030-y

CITATIONS READS

27 857

3 authors, including:

Ming Dai C.Q. Ru

Nanjing University of Aeronautics & Astronautics University of Alberta
68 PUBLICATIONS   636 CITATIONS    176 PUBLICATIONS   8,130 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Dislocations interacting with surfaces/interfaces in nanostructures View project

Inverse problems of holes/inclusions in linear elasticity View project

All content following this page was uploaded by Ming Dai on 14 November 2015.

The user has requested enhancement of the downloaded file.

Meccanica (2014) 49:2847–2859
DOI 10.1007/s11012-014-0030-y

Surface tension-induced stress concentration around

a nanosized hole of arbitrary shape in an elastic half-plane
Ming Dai • Cun-Fa Gao • C. Q. Ru

Received: 21 January 2014 / Accepted: 14 July 2014 / Published online: 29 July 2014
 Springer Science+Business Media Dordrecht 2014

Abstract This paper studies surface tension-induced hole and the free surface of the half-plane. On the other
stress concentration around a nanosized hole of hand, if the distance between the hole and the free
arbitrary shape inside an elastic half-plane. Of partic- surface is more than three times the hole size, the
ular interest is the maximum hoop stress on the hole’s effect of the free surface on the stress concentration
boundary with relation to the point of maximum around the hole is ignorable and the elastic half-plane
curvature and the distance between the hole and the can be treated approximately as an elastic whole plane.
free surface of the half-plane. The shape of the hole is
characterized by a conformal mapping which maps the Keywords Hole  Stress concentration  Surface
exterior of the hole onto the exterior of the unit circle tension  Hoop stress  Half-plane
in the image plane. On using the technique of
conformal mapping and analytic continuation, the
complex potentials of the half-plane are expressed in a
series form with unknown coefficients to be deter- 1 Introduction
mined by Fourier expansion method. Detailed numer-
ical results are shown for elliptical, triangular, square The analysis of deformation and stress in elastic solids
and rectangular holes. Two basic conclusions are that with inhomogeneities or holes is of great importance
the hoop stress increases with decreasing hole size and in materials science, solid-state physics and mechanics
the maximum hoop stress generally appears nearby of composites. For example, Muskhelishvili [1] stud-
but not exactly at the point of maximum curvature. In ied the elastic field of an infinite isotropic plane with
addition, it is shown that the hoop stress nearby the one elliptical hole, Savin [2] studied many important
point of maximum curvature on the hole’s boundary cases of stress concentration around holes, and
increases rapidly with decreasing distance between the Eshelby [3, 4] established the well-known Eshelby
formalism for elastic inclusion/inhomogeneity. As the
sizes of materials and structures down to the nano-
M. Dai  C.-F. Gao (&)
scale, however, surface tension and surface energy,
State Key Laboratory of Mechanics and Control of
Mechanical Structures, Nanjing University of Aeronautics which are usually negligible in macroscale, can come
and Astronautics, Nanjing, China to play a relevant role, and the material properties and
e-mail: cfgao@nuaa.edu.cn elastic field around interior holes or inhomogeneities
can be size-dependent due to increasing ratio of
C. Q. Ru
Department of Mechanical Engineering, University of surface to bulk volume and the influence of surface
Alberta, Edmonton, AB T6G 2G8, Canada tension and surface energy [5–7].

123
2848 Meccanica (2014) 49:2847–2859

To mention a few, Sharma et al. [8, 9] studied the Free surface

size-dependent elastic state of eigenstrained spherical y
inhomogeneities and obtained modified Eshelby’s x
tensor for spherical and cylindrical nanoinclusions.
Based on the Gurtin–Murdoch model [10–12], Tian
and Rajapakse [13, 14] investigated the two-dimen- S1
sional size-dependent elastic field of an infinite matrix
Γ1
with a nanoscale circular or elliptical inhomogeneity,
and Wang et al. [15, 16] studied the deformation and
stress concentration around a nanosized elliptical hole
with surface tension effect. Fang et al. [17] dealt with
two interacting coated nanowires with interface effect
under anti-plane shear waves, and Anjomshoa [18]
analyzed the buckling of embedded orthotropic circu- Fig. 1 A half-plane with an arbitrarily shaped hole with surface
tension
lar and elliptical micro/nano-plates based on nonlocal
elasticity theory. Shen et al. [19] investigated the
2 Problem descriptions and basic equations
imperfect interface effect on the stress field around an
elliptical inclusion in anti-plane shear, while Luo and
2.1 Problem descriptions
Gao [20] studied the stress field around an arbitrary-
shape inclusion by modeling the interface between
Shown in Fig. 1 is an elastic half-plane S1 of an
matrix and the inclusion as a thin coating. To the best
arbitrarily shaped hole bounded by the curve C1 in the
of our knowledge, almost all of previous works on
lower half-plane. The mechanical tractions, induced
elastic inhomogeneity/hole have been limited to
by surface tension T applied on the hole boundary C1,
elliptical inclusion and hole within an infinite elastic
can be described as
body. Actually, the non-elliptical shape of hole or
inhomogeneity as well as the presence of a traction- rn ¼ T  K; snt ¼ 0 ð1Þ
free surface near the hole or inhomogeneity could have where rn and snt are the normal and tangential
significant impact on the local deformation and stress tractions imposed on the hole boundary C1, T is the
distribution around the hole or inhomogeneity. There- constant surface tension, and K is point-wise curvature
fore, it is of great interest to study surface tension- along the boundary curve C1 (for the complicated
induced elastic field around a non-elliptical hole inside expression of the curvature K in terms of the mapping
an elastic half-plane. function which defines the hole shape C1, see Eq. (20)
In the present work, we present a series solution and Appendix).
to the surface tension-induced elastic field around
an arbitrarily shaped hole inside an elastic half- 2.2 Basic equations
plane. In Sect. 2, the present problem is described,
followed by the basic equations obtained with the The stresses (rx, ry, sxy) and displacements (u, v) in the
aid of analytic continuation techniques. In Sect. 3, elastic half-plane S1 can be given by two analytic
the complex potentials of the half-plane are functions, u(z) and w(z), as [1]
presented in a series form to be solved by means h i
of Fourier series expansions. In Sect. 4, results rx þ ry ¼ 2 u0 ðzÞ þ u0 ðzÞ ;
obtained by the present solution are first compared ry  rx þ 2isxy ¼ 2½zu00 ðzÞ þ w0 ðzÞ;
with some known solutions for the simpler case of
2Gðu þ ivÞ ¼ juðzÞ  zu0 ðzÞ  wðzÞ ð2Þ
circular and elliptical holes, respectively, and then
numerical examples are discussed in detail for where z = x ? iy, j = 3–4l for plane strain and
elliptical, triangular, square and rectangular holes, j = (3-l)/(1 ? l) for plane stress with G and l
respectively. Finally, the main results of the present representing the shear modulus and Poisson’s ratio.
work are summarized in Sect. 5. Assuming that no loads are applied at infinity, one has

123
Meccanica (2014) 49:2847–2859 2849

u0 ðzÞ ¼ oð1Þ; w0 ðzÞ ¼ oð1Þ; jzj ! þ1 ð3Þ entire z-plane with two holes bounded by C1 and C2,
but may be multi-valued. Thus, u(z) and w(z) in the
and the constant terms in u(z) and w(z), which do not domain S1 can be expressed in terms of f(z) by
influence the stress field, can be neglected.
In Fig. 2, S2 denotes the upper half-plane with a uðzÞ ¼ f ðzÞ; wðzÞ ¼ f ðzÞ  zf 0 ðzÞ; z 2 S1 ð7Þ
hole bounded by C2, while C1 and C2 are symmetrical
about the x-axis. Introduce a function defined in the In what follows, we shall determine the function
two half-planes S1 and S2, separately f(z), from which the functions u(z) and w(z) in the
 elastic half-plane S1 can be obtained through Eq. (7).
zu0 ðzÞ  wðzÞ z 2 S2 ;
f ðzÞ ¼ ð4Þ
uðzÞ z 2 S1 ;
which is holomorphic in S1 and S2, respectively. Since 3 Theoretical derivations
there is no traction on the free surface of the elastic
half-plane S1, one has 3.1 Series representations

ry  isxy ¼ u0 ðzÞ þ u0 ðzÞ þ zu00 ðzÞ þ w0 ðzÞ ¼ 0; Firstly, let X1 and X2 denote the exterior of C1 and
y ¼ 0 ð5Þ C2 in the whole z-plane, respectively. Introduce a
conformal mapping function of truncated form ([1],
which means [21])
0
ðÞ 0
ðþÞ 0 0 !
f ðzÞ  f ðzÞ ¼ u ðÞ ðzÞ þ u ðþÞ ðzÞ XM
j
00 0
ðþÞ
z ¼ xðn1 Þ ¼ R0 þ R n1 þ mj n1 ð8Þ
ðþÞ
þ zu ðzÞ þ w ðzÞ ¼ 0 ð6Þ j¼1

and therefore the derivative of f(z) defined by (4) is which maps X1 onto the exterior of the unit circle in
continuous across the x-axis, so f(z) is analytic in the the n1-plane, and the complex constants R0, R and mj
(j = 1, …, M) can be specified by the location, the size
and the shape of C1. Due to the symmetry of C1 and C2
about the x-axis, one can map X2 onto the exterior of
the unit circle in the n2-plane by another related
y mapping function given by
!
XM
j
z ¼ xðn2 Þ ¼ R0 þ R n2 þ mj n2 ð9Þ
j¼1

Γ2 As shown in the Appendix, Eq. (9) does map X2

S2 onto the exterior of the unit circle in the n2-plane. In
particular, the same z in the z-plane will be mapped to
two different points n1 and n2 in the n-plane, although
the two points n1 and n2 in the n-plane correspond to
x each other through the simultaneous Eqs. (8) and (9).
Since the entire z-plane with two holes bounded by
C1 and C2 can be interpreted as the intersection of X1
S1 and X2, without loss of the generality, the unknown
Γ1 function f(z) can be always written as sum of one
function analytical in X1 and another function analyt-
ical in X2. Thus, based on the two dependent mapping
functions (8) and (9), f(z) can be expressed by two
multi-valued functions g1(n1) and g2(n2) in the exterior
Fig. 2 An entire plane with two symmetrical holes of arbitrary of the unit circle in the n1-plane and n2-plane,
shape respectively, as [1]

123
2850 Meccanica (2014) 49:2847–2859

f ðzÞ ¼ g1 ðn1 Þ þ g2 ðn2 Þ Substituting Eqs. (7) and (13) into Eq. (12) yields
XN X
N þ2
¼ a0 ln n1 þ aj nj þ b ln n þ bj nj rn  isnt ¼ 2Re½f 0 ðzÞ
1 0 2 2
  r2 x0 ðrÞ
j¼1 j¼1
 zf 00 ðzÞ  f 0 ðzÞ  f 0 ðzÞ  zf 00 ðzÞ ;
ð10Þ x0 ðrÞ
z ¼ xðrÞ
where the correspondence between z and n1 or z and n2
ð14Þ
is defined by (8) or (9), ln n1 and ln n2 are the multi-
valued terms, and aj (j = 0, …, N) and bj (j = 0, …, where, on using Eq. (10) and the chain rule, the
N ? 2) are unknown complex coefficients. It follows relevant derivatives of f(z) can be rewritten as
from Eqs. (2), (7) and the conditions for single-valued
displacements that g01 ðrÞ g02 ðgÞ
f 0 ðzÞ ¼ þ
x0 ðrÞ x0 ðgÞ
ja0  b0 ¼ 0 ð11Þ P
N P
Nþ2
a0 r1  jaj rj1 b0 g1  jbj gj1
In what follows, the boundary conditions (1) on the j¼1 j¼1
hole will be used to determine all (2N ? 4) unknowns ¼ þ
x0 ðrÞ x0 ðgÞ
aj (j = 0, …, N) and bj (j = 0, …, N ? 2).
ð15Þ
3.2 Boundary conditions along the hole g0 1 ðgÞ g0 2 ðrÞ
f 0 ðzÞ ¼ þ
x0 ðgÞ x0 ðrÞ
By using the stress expressions in Eq. (2) together with
P
N P
Nþ2
stress transformation formula, the normal and tangen- a0 g1  jaj gj1 b0 r1  jbj rj1
j¼1 j¼1
tial tractions on the hole boundary C1 can be written as ¼ þ
x0 ðgÞ x0 ðrÞ
0 00 0 2ia
rn  isnt ¼ 2Re½u ðzÞ  ½zu ðzÞ þ w ðzÞe ;
ð16Þ
z ¼ xðrÞ ð12Þ

g001 ðrÞx0 ðrÞ  x00 ðrÞg01 ðrÞ g002 ðgÞx0 ðgÞ  x00 ðgÞg02 ðgÞ
f 00 ðzÞ ¼ 3
þ  3
ðx0 ðrÞÞ x0 ðgÞ
! !
2
P
N
j2 0 00 1
P
N
j1
a0 r þ jðj þ 1Þaj r x ðrÞ  x ðrÞ a0 r  jaj r
j¼1 j¼1
¼
0 ðrÞÞ3
ðx! !
P
Nþ2 P
Nþ2
b0 g2 þ jðj þ 1Þbj gj2 x0 ðgÞ  x00 ðgÞ b0 g1  jbj gj1
j¼1 j¼1
þ  3 ð17Þ
x0 ðgÞ

where r = eih represents the value of n1 on the unit where, in view of (8) and (9), g is related to r by
circle in the n1-plane and a is the angle between the
xðgÞ ¼ xðrÞ; jgj [ 1 ð18Þ
outer normal on C1 and the positive x-axis in the z-
plane, which is determined by According to the mapping function (9), Eq. (18) can
be rewritten as a (M ? 1) order equation of g, as
r2 x0 ðrÞ
e2ia ¼ ð13Þ follows
x0 ðrÞ

123
Meccanica (2014) 49:2847–2859 2851

  X
M 3.3 General solutions
RgMþ1 þ R0  xðrÞ gM þ Rmj gMj ¼ 0;
j¼1 In what follows, we employ Fourier series method to
jgj [ 1 ð19Þ solve Eq. (21) approximately. To this end, first,
expand both sides of Eq. (21) into truncated Fourier
from which we can get g by numerical computation for
series (power series of r) as
any given r. Here, among (M ? 1) roots of g obtained
by solving Eq. (19) numerically for an arbitrarily X
N X
N þ1
ð1Þ
N þ2 X
X N þ1
ð2Þ
given r, there exists only one g satisfying |g| [ 1, Cj;k rk aj þ Cj;k rk bj
j¼0 k¼N1 j¼0 k¼N1
because the mapping function (9) is bijective.
Next, the curvature K along C1 can be written in X
N X
Nþ1
ð1Þ
Nþ2 X
X Nþ1
ð2Þ
þ C j;k rk aj þ C j;k rk bj
terms of x(r) as [22] j¼0 k¼N1 j¼0 k¼N1
h i X
N X
N þ1 N þ2 X
X N þ1
ð3Þ ð4Þ
jx0 ðrÞj2 þRe rx0 ðrÞx00 ðrÞ þ Cj;k rk aj þ Cj;k rk bj
K¼ ð20Þ j¼0 k¼N1 j¼0 k¼N1
jx0 ðrÞj3
X
N X
Nþ1
ð5Þ
X
Nþ2 X
Nþ1
ð6Þ
and the detailed derivation is shown in the Appendix. þ Cj;k rk aj þ Cj;k rk bj
j¼0 k¼N1 j¼0 k¼N1
Thus, substituting Eqs. (14)–(17) together with
Eq. (20) into Eq. (1), one gets the stress boundary X
N þ1
ð7Þ
¼T Ck rk ð22Þ
condition on C1, as follows: k¼N1

" # 8h i 9
r 1 X
N
jrj1 1
g X
N þ2
jgj1 < xðrÞ  xðrÞ ½x0 ðrÞ  x00 ðrÞr=
2Re a0 þ aj þ 0 b0 þ bj þ a
x0 ðrÞ j¼1
x 0 ðrÞ
x ðgÞ 0
j¼1 x ðgÞ
: x0 ðrÞðx0 ðrÞÞ2 ; 0
8h i 9
XN < xðrÞ  xðrÞ ½jðj þ 1Þx0 ðrÞrj þ jx00 ðrÞrjþ1 =
þ a
j¼1
: x0 ðrÞðx0 ðrÞÞ2 ; j
8 h i 9
<r2 x0 ðrÞ xðrÞ  xðrÞ x0 ðgÞg2  x00 ðgÞg1 =
þ  3 b
: x0 ðrÞ x0 ðgÞ ; 0
8 h i 9
N þ2 <r2 x0 ðrÞ xðrÞ  xðrÞ jðj þ 1Þx0 ðgÞgj2 þ jx00 ðgÞgj1 =
X r2 x0 ðrÞg1
þ   b j þ a0
j¼1
: x0 ðrÞ x0 ðgÞ
3 ; x0 ðrÞx0 ðgÞ
XN
jr2 x0 ðrÞgj1 r X
N þ2
jrjþ1 r XN
jrjþ1 r2 x0 ðrÞg1
þ aj þ b0 þ bj þ a0 þ aj þ b0
0 0
x ðrÞx ðgÞ 0
x ðrÞ 0 0 0 x0 ðrÞx0 ðgÞ
j¼1 j¼1 x ðrÞ x ðrÞ j¼1 x ðrÞ
X
N þ2
jr2 x0 ðrÞgj1
þ bj
j¼1 x0 ðrÞx0 ðgÞ
h i
jx0 ðrÞj2 þRe rx0 ðrÞx00 ðrÞ
¼T ð21Þ
jx0 ðrÞj3

As explained below, the (2 N ? 4) unknowns aj

(j = 0, …, N) and bj (j = 0, …, N ? 2) can be
obtained from (21).

123
2852 Meccanica (2014) 49:2847–2859

where Z2p
ð6Þ 1 r
Z2p C0;k ¼ rk dh;
ð1Þ 1 r 1 2p x0 ðrÞ
C0;k ¼ rk dh; 0
2p x0 ðrÞ Z2p
0 ð6Þ 1 jrjþ1
Z2p Cj;k ¼ rk dh; j ¼ 1; . . .; N þ 2
ð1Þ 1 jrj1 k 2p x0 ðrÞ
Cj;k ¼ r dh; j ¼ 1; . . .; N 0
2p x0 ðrÞ
0 ð28Þ
ð23Þ h i
Z2p jx0 ðrÞj2 þRe rx0 ðrÞx00 ðrÞ
ð7Þ 1
Z2p Ck ¼ rk dh ð29Þ
ð2Þ 1 g1 k ð2Þ
2p jx0 ðrÞj3
C0;k ¼ r dh; Cj;k 0
2p x0 ðgÞ
0 Here, the definite integrals in Eqs. (23)–(29) are
Z2p j1 ð24Þ calculated by Gaussian quadrature. Then, letting the
1 jg
¼ rk dh; coefficients of rk (k = -N - 1, …, N ? 1) on both
2p x0 ðgÞ
0 sides of Eq. (22) be equal, respectively, one can get
j ¼ 1; . . .; N þ 2 (2 N ? 3) complex linear equations, as follows:

8h i 9
Z2p < xðrÞ  xðrÞ ½x0 ðrÞ  x00 ðrÞr
ð3Þ 1 r = k
C0;k ¼ þ r dh;
2p : x0 ðrÞðx0 ðrÞÞ2 x0 ðrÞ;
0 h i
Z2p xðrÞ  xðrÞ ½jðj þ 1Þx0 ðrÞrj þ jx00 ðrÞrjþ1  Z2p
ð3Þ 1 1 jrjþ1
Cj;k ¼ 2
rk dh þ rk dh; j ¼ 1; . . .; N
2p x0 ðrÞðx0 ðrÞÞ 2p x0 ðrÞ
0 0
ð25Þ
i h
Z2p r2 x0 ðrÞ xðrÞ  xðrÞ x0 ðgÞg2  x00 ðgÞg1  Z2p 2 0
ð4Þ 1 k 1 r x ðrÞg1 k
C0;k ¼  3 r dh þ r dh;
2p x0 ðrÞ x0 ðgÞ 2p x0 ðrÞx0 ðgÞ
0 0
h i
Z2p r2 x0 ðrÞ xðrÞ  xðrÞ jðj þ 1Þx0 ðgÞgj2 þ jx00 ðgÞgj1 
ð4Þ 1
Cj;k ¼  3 rk dh
2p x01 ðrÞ x0 ðgÞ ð26Þ
0
Z2p
1 jr2 x0 ðrÞgj1
þ rk dh;
2p x0 ðrÞx0 ðgÞ
0
j ¼ 1; . . .; N þ 2

Z2p N   Nþ2  
ð5Þ 1 r2 x0 ðrÞg1 X X
C0;k rk dh; ð1Þ ð3Þ
Cj;k þ Cj;k aj þ
ð2Þ ð4Þ
Cj;k þ Cj;k bj
2p x0 ðrÞx0 ðgÞ
0 j¼0 j¼0
Z2p N 
X  N þ2 
X 
ð5Þ 1 jr2 x0 ðrÞgj1 þ
ð1Þ ð5Þ
C j;k þ Cj;k aj þ
ð2Þ ð6Þ
C j;k þ Cj;k bj
Cj;k rk dh; j ¼ 1; . . .; N
2p x0 ðrÞx0 ðgÞ j¼0 j¼0
0
ð7Þ
ð27Þ ¼T Ck ; k ¼ N  1; . . .; N þ 1 ð30Þ

123
Meccanica (2014) 49:2847–2859 2853

-1

-2

d ⋅R -3
N=15
N=19
β R -4
N=20
-5

σhR/T
-6

-7

-8 d=0.1
-9

-10

-11
0.0 0.2 0.4 0.6 0.8 1.0
β /π (a)
-1.0 -1.0
N=5 N=3
-1.2
N=7 -1.1 N=5
-1.4 N=8 N=6
-1.2
-1.6 σhR/T
σhR/T

-1.3
-1.8

-2.0 -1.4

-2.2 -1.5

-2.4
d=0.5 d=1.0
-1.6
-2.6
-1.7
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
β /π (b) β /π (c)
Fig. 3 Convergence of the hoop stress around a circular hole

Finally, the (2 N ? 4) unknowns aj (j = 0,…, distance between the hole and the free surface of the
N) and bj (j = 0,…, N ? 2) are obtained through half-plane. Here, a ‘‘guess and check’’ method is used
solving the combination of Eqs. (11) and (30). to obtain a convergent result. Generally, in order to
achieve a satisfactory result, a larger number N is
usually required for a more complicated mapping (8)
4 Numerical examples or a smaller distance between the hole and the free
surface of the half-plane. In the following numerical
4.1 Verification of the present method examples, the hoop stresses obtained by the present
solution are all based on convergent solutions.
The convergence of the hoop stress around a circular Figure 5 shows the hoop stress around a circular
hole and an elliptical hole, with different distance hole with different distance between the circular hole
between the holes and the free surface of the half- and the free surface of the half-plane. It can be seen
plane, are shown in Figs. 3 and 4, respectively. In that the hoop stress obtained by our present solution,
Figs. 3 and 4, the convergence is demonstrated by the are in good agreement with those of Ref. [2] obtained
fact that the relative error between the hoop stresses by the curvilinear coordinate system. It is noticed that
corresponding to two adjacent N is less than 1 %. It is the minimum absolute value of hoop stress, which
found that for different hole shapes the required always appears at the upper and lower endpoints of the
number N corresponding to a convergent hoop stress circular hole, remains unchanged regardless of the
appears to be different and it increases with decreasing distance between the hole and the free surface.

123
2854 Meccanica (2014) 49:2847–2859

2
1 N=13
d ⋅a 0 N=15
-1 N=16
β 0.5a
-2
-3
1.5a

σha/T
-4
-5
-6
-7
-8 d=0.1
-9
-10
-11
0.0 0.2 0.4 0.6 0.8 1.0
β /π (a)
-0.6
-0.7

-1.1
-1.2
-1.5

-1.9
-1.8
-2.3
σha/T

σha/T

-2.7

-3.1 -2.4
N=9 N=8
-3.5 N=11 d=0.5 N=10 d=1.0
-3.9 N=12 -3.0 N=11
-4.3

-4.7 -3.6
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
β /π (b) β /π (c)
Fig. 4 Convergence of the hoop stress around an elliptical hole

Fig. 5 Hoop stresses 0

around a circular hole
-2

-4
present solution Ref. [2]
-6 d=0.05 d=0.05
d=0.1 d=0.1
-8
d=0.2 d=0.2
d=0.5
σhR/T

-10 d=0.5
d=1.0 d=1.0
-12

-14 d ⋅R

β R
-16

-18

-20

-22
0.0 0.2 0.4 0.6 0.8 1.0
β /π

123
Meccanica (2014) 49:2847–2859 2855

Fig. 6 Hoop stress around a -0.5

horizontal elliptical hole
-1.0

-1.5 present solution

m=0.6
-2.0 m=0.4
m=0.2

σha/T
-2.5

-3.0 FEM
m=0.6
-3.5 1.5a m=0.4
β (1 − m)a
(1 + m)a m=0.2
-4.0

-4.5

-5.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
β /π

0
2

0 -2
d=0.1
d=0.2
d=0.1
-2 -4 d=0.4
d=0.2
a/T

d=3.0
a/T

d=0.4
d
h

-4 d=2.5
h

-6
d d ⋅a
-6
d ⋅a
-8
1.5a
0.5a
-8 1.5a 0.5a

-10
-10

-12
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
/π (a) /π (b)
0

-2

-4

-6

-8
a/T

d=0.1
-10 d ⋅a d=0.2
h

-12 d=0.4
π4 d=2.8
-14 d
1.5a 0.5a

-16

-18

-20
0.0 0.4 0.8 1.2 1.6 2.0
/π (c)
Fig. 7 Hoop stress around elliptical holes with different orientations

123
2856 Meccanica (2014) 49:2847–2859

-2 -2

-4 -4
d=0.05 -6
-6
d=0.1 d=0.1
-8 -8
d=0.2 d=0.2
-10 d=1.5 -10
d=0.4
d -12
a/T

-12 d=1.5

a/T
-14 d
-14
h

h
-16
-16
d ⋅a -18 d ⋅a
-18
-20
-20 a
-22
-22
a -24
-24
-26
-26
-28
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
/π (a) /π (b)

Fig. 8 Hoop stress around triangular holes with different orientations

0
-1
-2
-2
-4
-3
d=0.1
-6 d=0.1
d=0.2 -4 d=0.2
d=0.4
-8 d=0.4
d=2.0 -5
a/T

d=3.0
a/T

d
-10 d
h

-6
h

-12
d ⋅a -7 d ⋅a

-14 -8 a
a
-16 -9

-18 -10

-20 -11
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
/π (a) /π (b)
Fig. 9 Hoop stress around a square hole

Shown in Fig. 6 is the hoop stress around a computing time (from a few seconds to a few tens of
horizontal elliptical hole with different ellipticity, seconds) is requested to obtain a satisfactory result.
with comparison to our finite element results. It is seen
from Fig. 6 that the hoop stress, obtained by our
present solution, is very close to those obtained by the 4.2 Further results on holes of arbitrary shape
finite element method. Actually, due to high stress
concentration around the hole, the results calculated For an elliptical hole, the mapping function (8) has an
by the finite element method may be sensitive to the exact form [1]

mesh quality. Therefore, in order to obtain a satisfac- a0 þ b0 a0  b0 1 ic

tory result by the finite element method, the finite z ¼ xðn1 Þ ¼ z0 þ n1 þ n1 e
2 2
element models of different mesh quality have been
ð31Þ
calculated until a satisfactory one is achieved. Con-
sequently, solving the present problem, of surface while for an equilateral triangular hole, a square hole
tension-induced stress concentration around a hole, by and a rectangular hole of a side ratio 3.24:1, the
the finite element method is very time-consuming. mapping function (8) can be approximately written,
With our present solution, however, only a little respectively, as [1, 2, 23]

123
Meccanica (2014) 49:2847–2859 2857

0
2
-2
0
d=0.1
-2 -4
d=0.2
d=0.1 d=0.4
-4 d=0.2 -6 d=3.0
d=0.4 d
-6 d=2.5 -8
a/T

a/T
d
-8
h

h
-10 d ⋅a
-10 d ⋅a
-12
-12 a
3.24a
3.24a -14
-14

-16 -16
a

-18 -18
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
/π (a) /π (b)
Fig. 10 Hoop stress around a rectangular hole
pffiffiffi

3a 1 2 ic points of (local) maximum curvature such as the

z ¼ xðn1 Þ ¼ z0 þ n1 þ n1 e ð32Þ endpoints of the major axis of the elliptical hole and
4 3
the vertices of the triangular, square and rectangular

3a 1 holes. It is seen from Figs. 7, 8, 9, and 10 that nearby the

z ¼ xðn1 Þ ¼ z0 þ n1  n3 eic ð33Þ endpoint or vertex which is most close to the free
5 6 1
surface, the hoop stress increases steeply with decreas-
z ¼ xðn1 Þ ing distance between the hole and the free surface. But

40a 1 1 3 3 5 ic it should be noticed that the maximum hoop stress

¼ z0 þ n1 þ n1  n  n e generally does not appear exactly at the point of
33 2 1 8 1 80 1
ð34Þ maximum curvature. Moreover, when the distance
between the hole and the free surface is, say, two or
Here, z0 and c in Eqs. (31)–(34) characterize the three times the size of the hole, the stress distribution
center and orientation of the hole, a0 and b0 in Eq. (31) around the hole in a half-plane is almost identical to that
are the length of the semimajor and semiminor axis of around the same hole but in a whole elastic plane.
the elliptical hole, and a in Eqs. (32)–(34) represents
the (short) side length of the triangular, square and
rectangular holes. 5 Conclusions
The hoop stresses around the holes corresponding
to the above mapping functions (31)–(34) are shown in Surface tension-caused stress concentration around a
Figs. 7, 8, 9, and 10, respectively. hole in an elastic half-plane is examined, with an
Note that in Figs. 7, 8, 9, and 10, for stress field around emphasis on the effects of hole shape and the free
the hole, the half-plane can be regarded as a ‘‘whole surface of the half-plane. With the aid of the
plane’’ when the distance d between the hole and the free techniques of analytic continuation and conformal
surface tends to infinity. In particular, the results corre- mapping, the complex potentials of the half-plane are
sponding to d = ?? are obtained by letting the Fourier expressed in a series form with some unknown
ð2Þ ð4Þ ð5Þ coefficients which can be determined by the Fourier
coefficients Cj;k , Cj;k and Cj;k in Eq. (30) be all zero.
It is shown obviously from Figs. 7, 8, 9, and 10 that expansion method. Numerical examples are given to
the hoop stress rh is inversely proportional to the verify the accuracy and reliability of the present
parameter a, that is to say, the hoop stress increases with solution and to show the stress concentration around
decreasing hole size. On each of the elliptical, triangu- holes of various non-elliptical shapes. Several con-
lar, square and rectangular holes, there exist several clusions can be drawn as follows:

123
2858 Meccanica (2014) 49:2847–2859

(1) The surface tension-induced hoop stress

which shows that the mapping function (9) is bijective
increases with decreasing hole size and becomes
from the exterior of the unit circle in the n2-plane to X2
significant especially for nano-sized holes.
in the z-plane.
(2) The maximum hoop stress on the hole usually
Then, the derivative of the conformal mapping
appears nearby but not exactly at the point of
function (8) should satisfy
maximum curvature. !
(3) The hoop stress nearby the point of maximum XM

curvature increases steeply when the distance 8jn1 j  1; x0 ðn1 Þ ¼ R 1  jmj nj1
1 6¼ 0
j¼1
between the point of maximum curvature and
the free surface decreases and approaches zero. ð37Þ
(4) For an interior hole, the half-plane can be which means
treated approximately as a whole plane when
!
the distance between the hole and the free   X
M
j1
surface is at least three times the hole size. 8jn2 j ¼ n2   1; x0 ðn2 Þ ¼ R 1  jmj n2 6¼ 0
j¼1

Acknowledgments Dai and Gao thank the support from the ð38Þ
National Natural Science Foundation of China (11232007) and a
Conjugating Eq. (38) leads to
Project Funded by the Priority Academic Program Development
of Jiangsu Higher Education Institutions(PAPD), and Ru thanks
!
X
M
the support from the Natural Science and Engineering Research 8jn2 j  1; x0 ðn2 Þ ¼ R 1  jmj nj1
2 6¼ 0
Council of Canada (NSERC-RGPIN204992). j¼1

ð39Þ
Appendix
Finally, one can conclude from Eqs. (36) and (39)
that Eq. (9) is a mapping from the exterior of the unit
Proof of the mapping Eq. (9)
circle in the n2-plane to X2 in the z-plane.
Firstly, since the conformal mapping function (8) is
bijective, there exist one-to-one corresponding z1 and Derivation of the curvature K
n1 satisfying
!
XM In the z-plane, the point-wise coordinates (x, y) on C1
j
z1 ¼ xðn1 Þ ¼ R0 þ R n1 þ m j n1 ; can be written in terms of x(r) as
j¼1

z1 2 X1 ; jn1 j  1 ð35Þ x ¼ Re½xðrÞ; y ¼ Im½xðrÞ; r ¼ eih ð40Þ

Let z2 and n2 be the conjugates of z1 and n1, The first and second order derivatives of x and
respectively. So according to the conjugate of Eq. (35), y with respect to h can be expressed as
the one-to-one corresponding z2 and n2 satisfy 8
! > dx
X M
j
< ¼ Re½irx0 ðrÞ ¼ Im½rx0 ðrÞ
z2 ¼ xðn2 Þ ¼ R0 þ R n2 þ m j n2 ; dh ð41Þ
j¼1 >
: dy ¼ Im½irx0 ðrÞ ¼ Re½rx0 ðrÞ
z2 2 X2 ; jn2 j  1 ð36Þ dh

8 2
>
> d x  
< 2 ¼ Im½irðx0 ðrÞ þ rx00 ðrÞÞ ¼ Re rx0 ðrÞ þ r2 x00 ðrÞ
dh ð42Þ
>
> d2 y  
: 2 ¼ Re½irðx0 ðrÞ þ rx00 ðrÞÞ ¼ Im rx0 ðrÞ þ r2 x00 ðrÞ
dh

123
Meccanica (2014) 49:2847–2859 2859

Finally, the signed curvature K can be given by the 10. Gurtin ME, Murdoch AI (1975) A continuum theory of
curvature formula as elastic material surfaces. Arch Ration Mech Anal
57:291–323
dx d 2 y dy d2 x 11. Gurtin ME, Murdoch AI (1978) Surface stress in solids. Int J
 Solids Struct 14:431–440
dh dh2 dh dh2
K ¼ "

2 #3=2
12. Gurtin ME, Weissmuller J, Larché F (1998) A general
dx 2 dy theory of curved deformable interfaces in solids at equi-
þ librium. Philos Mag A 78:1093–1109
dh dh 13. Tian L, Rajapakse RKND (2007) Analytical solution for
h i
size-dependent elastic field of a nanoscale circular inho-
jx0 ðrÞj2 þRe rx0 ðrÞx00 ðrÞ mogeneity. ASME J Appl Mech 74:568–574
¼ ð43Þ 14. Tian L, Rajapakse RKND (2007) Elastic field of an isotropic
jx0 ðrÞj3
matrix with a nanoscale elliptical inhomogeneity. Int J
Solids Struct 44:7988–8005
15. Wang GF, Wang TJ (2006) Deformation around a nanosized
elliptical hole with surface effect. Appl Phys Lett
89:161901–161903
References 16. Zeng XW, Wang GF, Wang TJ (2011) Erratum: ‘‘Defor-
mation around a nanosized elliptical hole with surface
1. Muskhelishvili NI (1975) Some basic problems of the effect’’ [Appl. Phys. Lett. 89, 161901 (2006)]. Appl Phys
mathematical theory of elasticity. Noordhoff, Groningen Lett 98:159901–159902
2. Savin GN (1961) Stress concentration around holes. Perg- 17. Fang XQ, Zhang LL, Liu JX (2013) Dynamic stress con-
amon Press, London centration around two interacting coated nanowires with
3. Eshelby JD (1957) The determination of the elastic field of surface/interface effect. Meccanica 48:287–296
an ellipsoidal inclusion and related problems. Proc R Soc 18. Anjomshoa A (2013) Application of Ritz functions in
Lond A 241:376–396 buckling analysis of embedded orthotropic circular and
4. Eshelby JD (1959) The elastic field outside an ellipsoidal elliptical micro/nano-plates based on nonlocal elasticity
inclusion. Proc R Soc Lond A 252:561–569 theory. Meccanica 48:1337–1353
5. Wong EW, Sheehan PE, Lieber CM (1997) Nanobeam 19. Shen H, Schiavone P, Ru CQ, Mioduchowski A (2000) An
mechanics: elasticity, strength, and toughness of nanorods elliptic inclusion with imperfect interface in anti-plane
and nanotubes. Science 277:1971–1975 shear. Int J Solids Struct 37:4557–4575
6. Miller RE, Shenoy VB (2000) Size-dependent elastic 20. Luo JC, Gao CF (2011) Stress field of a coated arbitrary
properties of nanosized structural elements. Nanotechnol- shape inclusion. Meccanica 46:1055–1071
ogy 11:139–147 21. Krylov VI, Kantorovitch LV (1958) Approximate methods
7. Shenoy VB (2002) Size-dependent rigidities of nanosized of higher analysis. Wiley Interscience, London
torsional elements. Int J Solids Struct 39:4039–4052 22. Radok JRM (1955) Problems of plane elasticity for rein-
8. Sharma P, Ganti S, Bhate N (2003) Effect of surfaces on the forced boundaries. ASME J Appl Mech 77:246–254
size-dependent elastic state of nano-inhomogeneities. Appl 23. Zou W, He Q, Huang M, Zheng Q (2010) Eshelby’s problem
Phys Lett 82:535–537 of non-elliptical inclusions. J Mech Phys Solids 58:346–372
9. Sharma P, Ganti S (2004) Size-dependent Eshelby’s tensor
for embedded nano-inclusions incorporating surface/inter-
face energies. ASME J Appl Mech 71:663–671

123

View publication stats