Jtam-A4.dvi


JOURNAL OF THEORETICAL

AND APPLIED MECHANICS

55, 4, pp. 1279-1283, Warsaw 2017
DOI: 10.15632/jtam-pl.55.4.1279

TORSION OF A CIRCULAR ANISOTROPIC BEAM WITH TWO LINEAR

CRACKS WEAKENED WITH A CIRCULAR CAVITY

Etimad Bayramoǧlu Eyvazov, Mecit Öge

Bartın University, Faculty of Engineering, Department of Mechanical Engineering, Bartın, Turkey

e-mail: mecitoge@bartin.edu.tr; mecitoge@yahoo.com

Various componentsmade of anisotropicmaterials (plast-mass, glassmaterial, etc.) havebe-
en widely used in the production ofmodernmechanisms andmachinery. Precise calculation
of these elements, constituting the design, holds great importance. In general, fracture and
distribution are essential issues in safety calculations. In this study, torsion of a beamwith
an S oblast having outer and inner constraints asL2 and L1 circles with R2 and R1 radii,
respectively, is investigated.

Keywords: anisotropicmedium, orthotropic beams, isotropic beams, affine connections, con-
formal mapping functions

1. Introduction

There is no study available in the literature a solution of the mentioned in the abstract cases,
since the contour L1 is void of the mapping function. In this study, the solution of the problem
is presented with numerical values.
Here, torsion of an area limited with an outer circle L2 with radius R2 and inner circle L1

with radius R1, having two linear cracks, is investigated. The coordinates of the end points of
these cracks are taken as±e. Volumetric forces are neglected.

Fig. 1. Anisotropic beam and its cross-section after affine transformation

The beam is twisted bymeans of a torsional moment applied to the edges (Fig. 1). Here, the
coordinate origin is taken as the center of cross-section. The beam is assumed to be made of a
homogeneous anisotropic material. At least one elastic plane of symmetry is available on each
point of the beam. In this case, all stresses except τxz and τyz are zero.
As known (Kosmodamianskii, 1976; Lethniskii, 1971; Kuliyev, 1991), solutions to torsion

problems related to orthotropic beams are found using the solutions to torsion problems of
beams with other cross-sections. In this case, affine connections are used



1280 E.B. Eyvazov,M. Öge

x1 =x and y1 =βy (1.1)

It should be noted that if the affine connection is assumed as below, then the beamwill not
be orthotropic

x1 =x+αy and y1 =βy (1.2)

With expression (1.1), none of the horizontal values in the beam cross-section changes (since
the horizontal axis does not change) as for the vertical values (i.e. those on Oy axis), they will
either increase or decrease depending on the coefficient β that characterizes anisotropy of the
beam.
Thereby, in order to evaluate stresses on orthotropic bars, firstly, the torsion problem of an

isotropic beamwithS1 cross-section (which is obtainedwith affine connection=x1 and y1 =βy)
should be solved.
According to the previous studies, τx1y1 and τy1z1 stresses are found from the following

equation

τx1z1 − iτy1z1 = i[2F
′(z1)−z1] (1.3)

where τx1z1 and τy1z1 are components of the tangential stresses on the cross-section S1, and
i – imaginary unit, F(z1) is the regular function on the cross-section S1, z1 = x1 + iy1 and
z1 =x1− iy1 are complex variables.F(z1) is calculated from boundary conditions, i.e., from the
states of equilibrium and the equation of deformations on the boundaries.
These boundaries can be written as follows

εx =0 εy =0 εz =0

γxy =0 γxz = a55τxz γyz = a44τyz

where a44 and a55 are elastic constants that characterize anisotropy of the material.
Equations (1.1) with affine connections, semi axes of L2 contour with radius R2 are trans-

formed into an ellipse a=x1 and b=βy1, on the other hand, the contourL1 with radiusR1 on
the Ox axis with two cracks, is transformed into an ellipse with two cracks (here it is assumed
that β 6=1).
If β < 1, then the linear values decline along the Oy axis, in the case of β > 1 the same

values increase.
F(z1) regular function within the enclosed area S1 can be evaluated using the below given

boundary conditions (Kosmodamianskii, 1976; Kuliyev, 1991, 2004; Sherman, 1992)

F(z1)−F(z1)= it1t1+Ck (1.4)

where t1 are affixes of the points on one of the contours of the cross-sectionS1.Ck is an arbitrary
constant.
Components of τx1z1 and τy1z1 tangential stresses on characteristic points of the cross-

-section S1 can be calculated by equation (1.4) (here, the end points of the cracks are also
included).
Afterwards, τxz and τyz can be calculated for an orthotropic beamby the following equation

(Kosmodamianskii, 1976; Lethniskii, 1971; Kuliyev, 1991)

τxz =βτx1z1 τyz = τy1z1

As indicated by these equations, τyz and τy1z1 stresses do not vary on isotropic and anisotropic
beams.Here, τxz tangential stress varies depending on the parameter β. It increases or decreases
depending on β (β=

√

a44/a55).



Torsion of a circular anisotropic beam with two linear cracks... 1281

F(z1) regular function can be expressed as follows for the contourL2 within the enclosed S1
area (Kosmodamianskii, 1976; Kuliyev, 1991, 2004; Sherman, 1992)

F(t2)= i
∞
∑

k=0

αk
(A2
t2

)k
+i

∞
∑

k=0

bk
( t2
A2

)k
on L2 (1.5)

where

αk =
k
∑

ν=k−2E(k/2)

∗

ανL(k−ν)/2 bk =
∞
∑

n=k

βka
(2)
(n−k)/2

(1.6)

F(z1) function can be defined as follows in the inner L1 contour (an ellipse with two linear
cracks) (Kosmodamianskii, 1976; Kuliyev, 1991; Sherman, 1992)

F(t1)= i
∞
∑

k=0

αkξ
−k
1 +i

∞
∑

k=1

H1(k)ξ
k
1 +i

∞
∑

k=0

H2(k)ξ
−k
1 on L1 (1.7)

The following notations are given in equation (1.7)

H1(k)=
∞
∑

ν=k

bν
(A1
A2

)ν
m
ν−k

2

1 C
ν−k

2
ν H2(k)=

∞
∑

ν=ε

∗

bν
(A1
A2

)ν
m
ν+k

2

1 C
ν+k

2
ν

ε′ = ε+
1

2
(k+ε) ε=0 ε=1

bk =
∞
∑

n=k

∗

βna(n−k)/2

(1.8)

The outer circle with contour L2 (with semi axes a2 = R2 and b2 = βR2) is mapped on the
circle with the unit diameter (equals to 1) using themapping function (Kosmodamianskii, 1976;
Kuliyev, 1991, 2004; Sherman, 1992)

t2 =A2
(

τ+
m2
τ

)

A2 =
R2+βR2
2

m2 =
R2−βR2
R2+βR2

(1.9)

The inner contour L1 is mapped onto the circle with the unit diameter (equals to 1) using the
mapping function (Kuliyev, 1991, 2004)

t1 =A1τ
∞
∑

n=0

τ−nΠn (1.10)

where

A1 =
R1+βR1
2

m1 =
R1−βR1
R1+βR1

Πn =
∞
∑

k=0

γk−1Tn−k Tn =
n
∑

ν=n−2E(n/2)

∗

m
n−ν

2

1 γ
n−ν

2

−1 l
n−ν
ν

The inverse functions of (1.9) and (1.10) mapping functions are as bellow (Kosmodamianskii,
1976; Kuliyev, 2004; Sherman, 1992)

ξ2 =
z2
A2

∞
∑

n=0

a(2)n

(A2
z2

)2n
ξ1 =χ(z) =

z1
A1

∞
∑

n=0

Ek
(A1
z1

)k
(1.11)



1282 E.B. Eyvazov,M. Öge

The coefficients of the serial elements of the analytical functions F1(z) and F2(z) can be fo-
und using proper boundary conditions given below: (Kosmodamianskii, 1976; Kuliyev, 1991;
Sherman, 1992)

F1(t1)+F1(t1)= it1+C1

F1(t2)+F1(t2)= it2+C2
(1.12)

where t1 and t2 variables are properly found from (1.9) and (1.10) equations.

As we place equations (1.6) and (1.7) into boundary condition (1.12), we obtain a linear
algebraic system depending on two unknown coefficients following some mathematical connec-
tions and remarks byKosmodamianskii (1976), Kuliyev (1991, 2004). Here, we proceedwith the
variable τ since ττ =1 on the unit circle)

αk +H1(k)+H2(k) =
∞
∑

n=k

Πn−kΠ on L1

V1(k)+V2(k)+V3(k)=A
2
2m2ε on L1

(1.13)

where V1(k), V2(k), V3(k) are respectively found from the following equations (Kuliyev, 1991,
2004)

V1(k)=
k
∑

ν=0

∗
(A1
A2

)k
C
k−ν

2

−ν m
k−ν

2

2 αν V2(k)=
∞
∑

ν=k

∗

bνC
k+ν

2
ν m

k+ν

2

2

V3(k)=
∞
∑

ν=k

∗

bνC
k−ν

2
ν m

k−ν

2

2

(1.14)

From the first terms of these equations, a system of equations is obtained. αk and βk are
coefficients that can be found using these equations.

This is presented with the following numerical example.

Cross-sectional dimensions of the beam are assumed in accordance with the following ratio
for numerical calculations.

1. In the case of β=1/2, the semi axes of the outer contour (curvilinear line, the circle with
radius R2) are transformed into the ellipse with semi axes a2 = R2; b2 = βR2, and the inner
contour (the one with two cracks and radiusR1) is transformed into the ellipse with two cracks.
The semi axes of such ellipses can be defined as a1 =R1; b1 = βR1. Accordingly, the problem
related with torsion of the beam depending on the parameter β is calculated using the torsion
problem of another beamwith a different cross section.

Tangential τxz and τyz stresses, given in Table 1, are found using equation (1.3).

Table 1.Tangential τxz and τyz stresses for choice 1

Choice 1 Points τxz/(µτa2) τyz/(µτa2)

β=1/2 z=0.65 – 24.58
a1 =2b1 z=0.70 – 13.29
a1 =0.25a2 z=0.75 – 8.02

z=1.00 – 2.92

z=0.5ia1 0.548 –
z=0.7ia1 0.141 –
z=1.0ia1 0.118 –



Torsion of a circular anisotropic beam with two linear cracks... 1283

2. In the case of β=1/2, R2 =0. The inner contour transforms into the linear crack with
length l = 2e; and the outer contour transforms into the ellipse with semi axes a2 = R2,
b2 = βR2. This way, the problem, the subject of the current study, is solved by means of the
solution to the torsion problem of the elliptical beamwith a central linear crack.
The values of τxz and τyz tangential stresses for choice 2 are given in Table 2.

Table 2.Values of τxz and τyz tangential stresses

Choice 1 Points τxz/(µτa2) τyz/(µτa2)

β=1/2 z=0.65 – 14.17
b1 =0 z=0.70 – 7.76
a1 =0.5a2 z=0.75 – 4.77

z=1.00 – 1.56

z=0.5ia1 −1.01 –
z=0.7ia1 −0.031 –
z=1.0ia1 −0.48 –

2. Conclusion

Calculations of torsion of orthotropic beams can be performed using calculations of isotropic
beamswith different cross-sections (the cross-sectionS is obtainedwith affine connectionx1 =x
and y1 = βy). Here, the linear values on the x axis do not vary with the varying parameter β),
the ones on the y axis increase or decrease in direct proportion with β.
The stresses on orthotropic beams can be calculated using the equations given below (τx1z1

and τy1z1 are known)

τxz =β
2ττx1y1 τyz =βττy1z1 (2.1)

References

1. Kosmodamianskii A.S., 1976, Stress State of Anisotropic Media with Holes and Cavities, Kiev-
-Donetsk Vysshaya Shkola, p.200

2. Kuliyev S.A., 1991,Two Dimensional Problems of Elasticity Theory, M. Stroyizdat, p.352

3. Kuliyev S.A., 2004, Conformally Mapping Functions of Complex Domains, Baku Azerneshr,
p.372

4. Lethniskii S.G., 1971,Torsion of Anisotropic and Inhomogeneous Bars, M. Nauka, p.200

5. Sherman D.I., 1992, Solution of planar elasticity problems in anisotropic media, Prikladnaya
Mathematika i Mekhanika, 180-106

Manuscript received November 10, 2016; accepted for print May 19, 2017