Upload
y-shindo
View
218
Download
2
Embed Size (px)
Citation preview
Acta Materialia 52 (2004) 4677–4684
www.actamat-journals.com
Tensile testing and analysis of ferromagnetic elastic strip witha central crack in a uniform magnetic field
Y. Shindo *, D. Sekiya, F. Narita, K. Hohiguchi
Department of Materials Processing, Graduate School of Engineering, Tohoku University, Aoba-yama 02, Sendai 980-8579, Japan
Received 28 April 2004; received in revised form 16 June 2004; accepted 16 June 2004
Available online 17 July 2004
Abstract
The effect of magnetic fields on the fracture mechanics parameters such as the stress intensity factor, energy density, etc., is
discussed by analyzing the plane strain and plane stress problems of a soft ferromagnetic strip with a central crack under a uniform
magnetic field. The problem of an infinitely long soft ferromagnetic elastic strip with a central crack is formulated by means of
integral transforms and reduced to the solution of a Fredholm integral equation of the second kind. Numerical values on the
fracture mechanics parameters are obtained. Tensile tests are also conducted on center-cracked soft magnetic plate with strain gage
technique, and the numerical predictions for the plane stress case are compared with the test results. Agreement between theory and
experiment is fair.
� 2004 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
Keywords: Theory and modeling; Tension test; Soft ferromagnetic materials; Fracture; Magneto-elastic interactions
1. Introduction
With the increasing applications of ferromagnetic
materials as fusion reactors, MHD devices, magneti-
cally levitated vehicles, etc., the theme of magneto-
elastic interactions has been addressed in the recent
past. If a ferromagnetic material is used in a magnetic
field, the combination of mechanical and magnetic
forces could produce elevated stresses and strains. Theferromagnetic materials may be degraded in such a
stress level. The strength of the ferromagnetic materials
is also weakened by the presence of defects such as voids
and cracks. It is therefore important to understand the
degradation phenomena of the ferromagnetic materials.
The fracture behavior of the ferromagnetic materials is
thus of much recent interest. The influence of the
magnetic field on the singular stress distributions nearthe crack tip in an infinite body was shown by Shindo
[1,2] based on a linear theory for the soft ferromagnetic
* Corresponding author. Tel./fax: +81-22-217-7341.
E-mail address: [email protected] (Y. Shindo).
1359-6454/$30.00 � 2004 Acta Materialia Inc. Published by Elsevier Ltd. A
doi:10.1016/j.actamat.2004.06.029
elastic materials of multidomain structure [3]. Consid-ered by Shindo [4] is the problem of determining the
stress intensity factors in an infinitely long strip of a soft
ferromagnetic elastic material with a crack parallel to
the edges of the strip.
This paper applies a linear theory for the ferromag-
netic elastic materials of multidomain structure [3] to the
problem of determining the distribution of stress in a
cracked soft ferromagnetic strip permeated by a uniformmagnetostatic field normal to the crack surface. The
plane strain and plane stress problems are considered for
the infinitely long strip containing a central crack nor-
mal to the edges of the strip. The method of solution
involves the use of Fourier transforms to reduce the
mixed boundary value problem to two simultaneous
dual integral equations. The solution is then given in
terms of a single Fredholm integral equation of thesecond kind. Numerical values on the fracture me-
chanics parameters such as stress intensity factor, energy
density, etc. are obtained. Tensile tests are also con-
ducted on soft magnetic plate with a central crack to
obtain the values of the fracture mechanics parameters,
ll rights reserved.
4678 Y. Shindo et al. / Acta Materialia 52 (2004) 4677–4684
and the numerical results for the plane stress are com-
pared with the experimental values.
2. Problem statement and basic equations
Consider a soft ferromagnetic isotropic linear elastic
strip of width 2h and length 2lðl ! 1Þ which contains a
central crack of length 2a aligned with its plane normal
to the free edges as shown in Fig. 1. A rectangular
Cartesian coordinate system ðx; y; zÞ is attached to be the
center of the crack for reference purposes. The x-axis isdirected along the line of the crack and y-axis along thedirection of the perpendicular bisector of the crack. The
edges of the soft ferromagnetic elastic strip are therefore
the lines with equations x ¼ �h, while the crack occupies
the segment �a < x < a, y ¼ 0. We consider a uniform
normal stress, ryy ¼ r0, applied with a uniform magnetic
field of magnetic induction B0y ¼ B0. Only the first
quadrant with appropriate boundary conditions needs
to be analyzed owing to symmetry.All magnetic quantities are divided into two parts,
those in the rigid body state and those in the perturba-
tion state as follows:
B ¼ B0 þ b;
M ¼ M0 þm;
H ¼ H0 þ h;
ð1Þ
where B, M and H are the magnetic induction, magne-
tization and magnetic intensity vectors, respectively. The
first parts, which are indicated by the subscript 0, are
magnetic quantities in the underformed body. The sec-
ond parts, which are represented by lower case letters,
are corrections to account for the additional changes in
140
mm
2a
2h
Weights
Straingage
P
Crack
B0
B0
Cryocooler-cooledsuperconducting magnet
xy
o
Fig. 1. Testing set-up.
magnetic quantities due to deformations. The magneto-
elastic solution for the rigid body state is
Bec0y ¼ B0; Hec
0y ¼B0
l0
; Mec0y ¼ 0;
06 x < a; y ¼ 0; ð2Þ
Be0y ¼
B0
lr
; He0y ¼
B0
l0lr
; Me0y ¼ 0;
h < x < 1; 0 < y < 1; ð3Þ
B0y ¼ B0; H0y ¼B0
l0lr
; M0y ¼vB0
l0lr
;
06 x6 h; 0 < y < 1; ð4Þ
where B0y , H0y and M0y are, respectively, the y-compo-
nents of B0, H0 and M0. The superscript eðh < x < 1;0 < y < 1Þ and ecð06 x < a; y ¼ 0Þ indicate values
outside the strip. Note that l0 ¼ 4p� 10�7 N/A2 (H/m;
H: Henry) is the magnetic permeability of the vacuum,
lr ¼ 1þ v is the specific magnetic permeability, and v is
the magnetic susceptibility.The effect of the magnetization as induced by the
deformation becomes important for a cracked soft fer-
romagnetic solid in a magnetic field normal to the crack
surface. In this case, the body force of the type
l0M � rH must be considered on account of the sharp
gradient of magnetic field near the crack. For the mag-
netic field in the solid, it is assumed that:
hx;x þ hy;y ¼ 0; ð5Þ
hx;y � hy;x ¼ 0 ð6Þand
bx ¼ l0lrhx;by ¼ l0lrhy ;
�ð7Þ
mx ¼ vhx;my ¼ vhy ;
�ð8Þ
where a comma denotes partial differentiation with re-spect to the coordinate, and ðhx; hyÞ, ðbx; byÞ and ðmx;myÞare the x and y-components of h, b and m. Eqs. (5) and
(6) for the perturbed state are satisfied by introducing a
magnetic potential / such that:
hx ¼ /;x; hy ¼ /;y ; ð9Þ
/;xx þ /;yy ¼ 0: ð10Þ
Using a dipole model for the magnetization, we ob-
tain the equations of equilibrium as
txx;x þ tyx;y þvB0/;xy
lr¼ 0;
txy;x þ tyy;y þvB0/;yy ¼ 0:
)ð11Þ
lr
Y. Shindo et al. / Acta Materialia 52 (2004) 4677–4684 4679
The components of magnetoelastic stresses txx; tyy ; txy ¼tyx and Maxwell stresses tMxx ; t
Myy ; t
Mxy ¼ tMyx are also obtained
as:
txx ¼ rxx;
tyy ¼ ryy þvB2
0
l0l2rþ 2vB0/;y
lr;
txy ¼ tyx ¼ rxy þ vB0/;x
lr;
9>=>; ð12Þ
rxx ¼ 2lux;x þ jðux;x þ uy;yÞ;ryy ¼ 2luy;y þ jðux;x þ uy;yÞ;rxy ¼ ryx ¼ lðux;y þ uy;xÞ;
9=; ð13Þ
tMxx ¼ � B0/;y
lr� B2
0
2l0l2r;
tMyy ¼ ð1þ2vÞB0/;y
lrþ ð1þ2vÞB2
0
2l0l2r;
tMxy ¼ tMyx ¼ B0/;x;
9>>=>>; ð14Þ
where rxx, ryy , rxy ¼ ryx are the elastic stress compo-
nents, ux and uy are the displacement components, j ¼ kfor plane strain, and j ¼ 2kl=ðkþ 2lÞ for plane stress,
and k ¼ 2Gm=ð1� 2mÞ and l ¼ G are the Lam�e constantswith G ¼ E=2ð1þ mÞ being the modulus of rigidity.
Making use of Eq. (12), the two Eqs. (11) become:
ux;xx þ ux;yy þjl
�þ 1
�ðux;x þ uy;yÞ;x þ
2vB0/;xy
llr
¼ 0;
ð15Þ
uy;xx þ uy;yy þjl
�þ 1
�ðux;x þ uy;yÞ;y þ
2vB0/;yy
llr
¼ 0:
ð16ÞThe mixed boundary conditions in the perturbation
state may be expressed as follows:
hecx ðx; 0Þ � hxðx; 0Þ ¼ �ðvB0=l0lrÞuy;xðx; 0Þ ð06 x < aÞ;/ðx; 0Þ ¼ 0 ða6 x6 hÞ; ð17Þ
becy ðx; 0Þ � byðx; 0Þ ¼ 0 ð06 x < aÞ;/ecðx; 0Þ ¼ 0 ð06 x < aÞ;
ð18Þ
ryxðx; 0Þ ¼ �ðvB0=lrÞhxðx; 0Þ ð06 x6 hÞ; ð19Þ
ryyðx; 0Þ ¼ ðvB0=lrÞ ðv�
� 2Þhyðx; 0Þ � B0=l0lr
�ð06 x < aÞ;
uyðx; 0Þ ¼ 0 ða6 x6 hÞ;ð20Þ
heyðh; yÞ � hyðh; yÞ ¼ 0 ðy < 1Þ; ð21Þ
hexðh; yÞ � lrhxðh; yÞ ¼ �ðvB0=l0lrÞux;yðh; yÞ ðy < 1Þ;ð22Þ
rxxðh; yÞ ¼ 0 ðy < 1Þ; ð23Þ
rxyðh; yÞ ¼ 0 ðy < 1Þ: ð24Þ
3. Solution procedure
Using now the Fourier transform, the components of
displacements and magnetic potential may be obtained
as:
ux ¼2
p
Z 1
0
AðaÞ�
þ y�
� jþ 3ljþ l
1
a
�BðaÞ
þ 2vB0
lrðjþ lÞ aðaÞ�e�ay sinðaxÞda
� 2
p
Z 1
0
CðaÞ sinhðaxÞ�
þ x coshðaxÞ�
� jþ 3ljþ l
sinhðaxÞ 1a
�DðaÞ
þ 2vB0
lrðjþ lÞ bðaÞ sinhðaxÞ�cosðayÞda� a0x;
ð25Þ
uy ¼2
p
Z 1
0
AðaÞf þ BðaÞyge�ay cosðaxÞda
þ 2
p
Z 1
0
CðaÞcoshðaxÞf þDðaÞx sinhðaxÞg sinðayÞda
þ b0y; ð26Þ
/ ¼ 2
p
Z 1
0
aðaÞe�ay cosðaxÞda
þ 2
p
Z 1
0
bðaÞ coshðaxÞ sinðayÞda; ð27Þ
where AðaÞ, BðaÞ, CðaÞ, DðaÞ, aðaÞ and bðaÞ are the un-
known functions to be solved, and a0, b0 are the real
constants, which will be determined from the far-field
loading conditions. Applying the Fourier transform to
Eq. (10) yields:
/ec ¼ 2
p
Z 1
0
aeðaÞsinhðayÞcosðaxÞda; y¼ 0; 06x< a;
ð28Þ
/e ¼ 2
p
Z 1
0
beðaÞe�ax sinðayÞda; h < x < 1; ð29Þ
where aeðaÞ and beðaÞ are also unknowns. The magnetic
fields can be obtained by making use of Eqs. (9) and(27). The magnetic fields in the void inside the crack and
outside the strip can also be obtained from Eqs. (9), (28)
and (29).
By applying the far-field loading conditions, the
constants a0 and b0 are obtained as:
a0 ¼j
4lðjþ lÞ r0
�� vB2
0
l0l2r
�;
b0 ¼jþ 2l
4lðjþ lÞ r0
�� vB2
0
l0l2r
�:
ð30Þ
4680 Y. Shindo et al. / Acta Materialia 52 (2004) 4677–4684
The boundary conditions of Eqs. (21)–(24) lead to the
following relations between unknown functions:
vB0
l0lr
asinhðahÞCðaÞþ vB0
l0lr
ahcoshðahÞ�
�jþ3ljþl
sinhðahÞ�DðaÞ
� lr
�� 2v2B2
0
l0l2r ðjþlÞ
�asinhðahÞbðaÞ�ae�ahbeðaÞ¼ f1ðaÞ;
ð31Þ
a coshðahÞbðaÞ�
� e�ahbeðaÞ�¼ f2ðaÞ; ð32Þ
a sinhðahÞCðaÞ þ ah coshðahÞ�
� ljþ l
sinhðahÞ�DðaÞ
þ vB0ðjþ 3lÞ2llrðjþ lÞ a sinhðahÞbðaÞ ¼ f3ðaÞ; ð33Þ
a coshðahÞCðaÞ þ ah sinhðahÞ�
� jþ 2ljþ l
coshðahÞ�DðaÞ
þ vB0ðjþ 2lÞllrðjþ lÞ a coshðahÞbðaÞ ¼ f4ðaÞ; ð34Þ
where fiðaÞ ði ¼ 1; 2; 3; 4Þ are given in Appendix A.
Application of the mixed boundary conditions in Eqs.
(17) and (20) gives rise to two simultaneous dual integral
equations:Z 1
0
a aðaÞ�
� vB0
l0lr
AðaÞ�sinðaxÞda ¼ 0; 06 x < a;
Z 1
0
aðaÞ cosðaxÞda ¼ 0; a6 x6 h;
Z 1
0
a AðaÞ�
�vB0ðjþ2lÞ2llrðjþlÞ v
�� ljþ2l
�aðaÞ
�cosðaxÞda
�jþ2ljþl
Z 1
0
E1ðahÞcoshðaxÞf þE2ðahÞaxsinhðaxÞgda
¼ pðjþ2lÞ4lðjþlÞr0; 06x< a;
Z 1
0
AðaÞ cosðaxÞda ¼ 0; a6 x6 h; ð35Þ
where EiðahÞ ði ¼ 1; 2Þ are given in Appendix B.
The solution of two simultaneous dual integral
equations (35) may be obtained by using a new functionUðnÞ and the result is
AðaÞ ¼ p2a2
Z 1
0
n1=2UðnÞJ0ðaanÞdn: ð36Þ
The function UðnÞ is governed by the following Fred-
holm integral equation of the second kind:
UðnÞ þZ 1
0
UðgÞKðn; gÞdg ¼ jþ 2l2lðjþ lÞy0
r0n1=2; ð37Þ
y0 ¼ 1� jþ 2l2ðjþ lÞ v
�� ljþ 2l
�vbclr
� �2
; ð38Þ
where b2c ¼ B20=ll0. The kernel function Kðn; gÞ is
Kðn; gÞ ¼ ðngÞ1=2 a2ðjþ 2lÞy0ðjþ lÞ
Z 1
0
ae�ahI0ðaagÞ
� L1ðahÞI0ðaanÞf þ aanL2ðahÞI1ðaanÞgda;ð39Þ
where LiðahÞ ði ¼ 1; 2Þ are given in Appendix C, and I0ðÞand I1ðÞ are the zero and first order modified Bessel
functions of the first kind, respectively.
The displacement components ux, uy and magnetic
potential / near the crack tip are:
ux � kh12l
r2
1=2 1a1cos h
2
2ljþl þ
vbclr
� �2� ��
þ 2þ jþ3ljþl
vbclr
� �2�sin2 h
2
� �;
uy � kh12l
r2
1=2 1a1sin h
2
2ðjþ2lÞjþl
h� 2þ jþ3l
jþlvbclr
� �2�cos2 h
2
� �;
9>>>>>>>>>=>>>>>>>>>;
ð40Þ
/ � kh1a1
r2
� �1=2 vB0ðjþ 2lÞll0lrðjþ lÞ sin
h2
ð41Þ
and the singular parts of the strains, elastic stresses,
magnetoelastic stresses and Maxwell stresses in the
neighborhood of the crack tip are:
exx � kh1a1E0ð2rÞ1=2
cos h2
ð1� m0Þ 1þ 2jþ3l2ðjþlÞ
vbclr
� �2� ��
�ð1þ m0Þ 1þ jþ3l2ðjþlÞ
vbclr
� �2�sin h
2sin 3h
2
� �;
eyy � kh1a1E0ð2rÞ1=2
cos h2
ð1� m0Þ 1þ 2jþ3l2ðjþlÞ
vbclr
� �2� ��
þ ð1þ m0Þ 1þ jþ3l2ðjþlÞ
vbclr
� �2� �
sin h2sin 3h
2
�;
exy � kh12a1Gð2rÞ1=2
sin h2
jþ2l2ðjþlÞ
vbclr
� �2�
� 1þ jþ3l2ðjþlÞ
vbclr
� �2� �
cos h2cos 3h
2
�;
9>>>>>>>>>>>>>>>>>>>=>>>>>>>>>>>>>>>>>>>;
ð42Þ
rxx � kh1a1ð2rÞ1=2
cos h2
1þ 2jþ3l2ðjþlÞ
vbclr
� �2� ��
� 1þ jþ3l2ðjþlÞ
vbclr
� �2� �
sin h2sin 3h
2
�;
ryy � kh1a1ð2rÞ1=2
cos h2
1þ 2jþ3l2ðjþlÞ
vbclr
� �2� ��
þ 1þ jþ3l2ðjþlÞ
vbclr
� �2� �
sin h2sin 3h
2
�;
rxy � kh1a1ð2rÞ1=2
sin h2
jþ2l2ðjþlÞ
h� 1þ jþ3l
2ðjþlÞvbclr
� �2� �
cos h2cos 3h
2
�;
9>>>>>>>>>>>>>>>>>>=>>>>>>>>>>>>>>>>>>;
ð43Þ
Y. Shindo et al. / Acta Materialia 52 (2004) 4677–4684 4681
txx � kh1a1ð2rÞ1=2
cos h2
1þ 2jþ3l2ðjþlÞ
vbclr
� �2� ��
� 1þ jþ3l2ðjþlÞ
vbclr
� �2� �
sin h2sin 3h
2
�;
tyy � kh1a1ð2rÞ1=2
cos h2
1þ 4jþ7l2ðjþlÞ
vbclr
� �2� ��
þ 1þ jþ3l2ðjþlÞ
vbclr
� �2� �
sin h2sin 3h
2
�;
txy � kh1a1ð2rÞ1=2
sin h2
jþ2ljþl
vbclr
� �2�
� 1þ jþ3l2ðjþlÞ
vbclr
� �2� �
cos h2cos 3h
2
�;
9>>>>>>>>>>>>>>>>>>>=>>>>>>>>>>>>>>>>>>>;
ð44Þ
tMxx � kh1a1ð2rÞ1=2
vðjþ2lÞ2ðjþlÞ
bclr
� �2
cos h2;
tMyy � kh1a1ð2rÞ1=2
vð1þ2vÞðjþ2lÞ2ðjþlÞ
bclr
� �2
cos h2;
tMxy �kh1
a1ð2rÞ1=2vðjþ2lÞ2ðjþlÞ
bclr
� �2
sin h2;
9>>>>=>>>>;
ð45Þ
where m0 ¼ m=ð1� mÞ, E0 ¼ E=ð1� m2Þ for plane strain,
and m0 ¼ m, E0 ¼ E for plane stress, and a1 is given by
a1 ¼ 1þ v2ðjþ lÞ ðjf þ 2lÞ þ vð2jþ 5lÞg bc
lr
� �2
:
ð46ÞThe polar coordinates r and h are defined as:
r ¼ ðxn
� aÞ2 þ y2o1=2
; ð47Þ
h ¼ tan�1 yx� a
� �: ð48Þ
The stress intensity factor kh1 is obtained as
kh1 ¼ limx!aþ
2ðxf � aÞg1=2tcyyðx; 0Þ ¼a1y0
r0a1=2Uð1Þ; ð49Þ
where
tcyyðx; 0Þ ¼ tyyðx; 0Þ þ tMyy ðx; 0Þ: ð50Þ
The energy density is expressible in the form
S ¼ r2
tcxxexx�
þ tcxyexy þ tcyxeyx þ tcyyeyy�
ð51Þ
and hence
S ¼ ahk2h1 ð52Þin which the coefficient ah depends on the angle h. For acrack under mode I loading, fracture will always occur
in the normal plane (h ¼ 0). The coefficient ah for h ¼ 0
is given by
ah ¼1� m0
2a21E0 1
(þ 2jþ 3l2ðjþ lÞ
vbclr
� �2)
� 1
"þ v2ðjþ lÞ jf þ 2lþ vð4jþ 7lÞg bc
lr
� �2#:
ð53Þ
4. Experimental procedure
Tensile tests were conducted on nickel–iron soft
magnetic materials with center-cracked plate specimen
geometry in the 100 mm diameter bore of a 10 T (T:Tesla) cryocooler-cooled superconducting magnet at
room temperature (Fig. 1). Three soft magnetic mate-
rials (TMC-V: Young’s modulus E ¼ 182 GPa, Pois-
son’s ratio m ¼ 0:146, specific magnetic permeability
lr ¼ 27,900; TMH-B: E ¼ 203 GPa, m ¼ 0:279, lr ¼10,690; TMB: E ¼ 146 GPa, m ¼ 0:228, lr ¼ 9030) de-
veloped by NEC/Tokin Co. Ltd. were selected. The
specimens are 140 mm length and 1 mm thickness. Thecrack length was varied (2a ¼ 10; 15; 20 mm) while
keeping the specimen width fixed at 2h ¼ 40 mm. Ini-
tial through-the-thickness notches were machined using
electro-discharge machining. The specimens were fa-
tigue precracked and then annealed to obtain the op-
timum magnetic properties. A simple strain gage
method is very suitable to determine the magnetic
stress intensity factor [5]. A five element strip gage(KFG-1-120-D19-16N10C2 from Kyowa Electronic
Instruments Co. Ltd., Japan) was installed along the
90� line and the center point of the element closest to
the crack tip was 2 mm. The strain sensors have an
active length of 1 mm.
Tensile load and a magnetic field were simultaneous
applied to center-cracked plate specimens. A supercon-
ducting magnet was used to create a static uniformmagnetic field of magnetic induction B0 normal to the
crack surface. The specimens were loaded by P ¼ 29:4 Nload that consisted of weights. For each specimen size
five tests were performed. The strains were recorded as a
function of magnetic field.
For the plane stress case, the strain eyy near the cracktip given by Eq. (42) can be rewritten in the form
Eeyy ¼kh1
ð2rÞ1=2v
2ð1þ mÞvþ 2mþð5� mÞvf g vbclr
� �2
� cosh2
ð1"
� mÞ 2ð1(
þ mÞþð3þ mÞ vbclr
� �2)
þð1þ mÞ 2ð1(
þ mÞþð3� mÞ vbclr
� �2)sin
h2sin
3h2
#
þA0þOðr1=2Þ; ð54Þ
where A0 is the unknown coefficient. If all the magnetic
field quantities are made to vanish, then Eq. (54) reduces
to the strain near the crack tip in an elastic plane body
[6]. Setting h ¼ p=2 gives
c0Eeyyr1=2 ¼ kh1 � 2c0A0r1=2 þ � � � ; ð55Þ
where
4682 Y. Shindo et al. / Acta Materialia 52 (2004) 4677–4684
c0 ¼4
v
2ð1þ mÞvþ 2mþ ð5� mÞvf g vbclr
� �2
2ð3þ 2m� m2Þ þ ð9� 2m� 3m2Þ vbclr
� �2p1=2: ð56Þ
From Eq. (55), a plot of c0Eeyyr1=2 versus r1=2 is linear forsmall values of r and the intercept at r ¼ 0, at the crack
tip, gives the stress intensity factor kh1.
Fig. 3. Stress intensity factor versus magnetic field (h=a ¼ 1:5, 2.0, 5.0).
Fig. 4. Stress intensity factor versus magnetic field (lr ¼ 100, 1000,
10,000).
5. Results and discussion
The determination of the stress intensity factor re-quires the solution of the function of UðnÞ. The solutionof the Fredholm integral equation of the second kind
(37) governing UðnÞ has been computed numerically by
the use of Gaussian quadrature formulas. Once this is
done, kh1 and S can be found from Eqs. (49) and (51).
Fig. 2 provides the normalized stress intensity factor
kh1=r0a1=2 as a function of h=a for m ¼ 0:25, lr ¼ 10,000
and the normalized magnetic field bc ¼ 0, 0.0032, 0.0047obtained from the plane stress and plane strain analyses.
For l ¼ 80 GPa, bc ¼ 0:0032 and 0.0047 correspond to
the magnetic induction of B0 ¼ 1:0 and 1.5 T, respec-
tively. The dashed curves obtained for bc ¼ 0 coincide
with the purely elastic plane stress and plane strain
cases. Comparing the results of these two cases for
bc ¼ 0, little difference is observed (two dashed curves
approximately overlap). The nomarlized stress intensityfactor tends to infinity as h=a ! 1, and decreases slowly
as h=a increases and tends to the result of the infinite
solid as h=a ! 1. Applying the magnetic field increases
the stress intensity factor. The values of kh1=r0a1=2 for
h=a ! 1 are found to be kh1=r0a1=2 ¼ 1:089, 1.215
(bc ¼ 0:0032, 0.0047) for plane stress and kh1=r0a1=2 ¼1:083, 1.199 (bc ¼ 0:0032, 0.0047) for plane strain. Fig. 3exhibits the normalized stress intensity factor kh1=r0a1=2
as a function of bc for m ¼ 0:25, lr ¼ 10,000 and h=a ¼1:5, 2.0, 5.0 obtained from the plane stress analysis. The
magnetic field effect can increase the values of the stress
intensity factor and depends on h=a. Fig. 4 shows the
variation of the normalized stress intensity factor
Fig. 2. Stress intensity factor versus strip-width to crack-length ratio
(bc ¼ 0, 0.0032, 0.0047).
kh1=r0a1=2 against bc for m ¼ 0:25, h=a ¼ 2:0 andlr ¼ 1000, 5000, 10,000 in the plane stress case. The
effect of the magnetic field on the stress intensity factor
is more pronounced with increasing the specific mag-
netic permeability. Fig. 5 shows the energy density S for
h ¼ 0 under different bc for m ¼ 0:25, lr ¼ 10,000 and
h=a ¼ 1:5, 2.0, 5.0, where S has been normalized by the
energy density S0 of infinite ferromagnetic elastic solid
for bc ¼ 0. All the curves increase with increasing bc.The energy density tends to increase with decreasing
h=a.
Fig. 5. Energy density factor versus magnetic field.
Fig. 6. Stress intensity factor versus strip-width to crack-length ratio
(TMC-V).
Fig. 7. Stress intensity factor versus strip-width to crack-length ratio
(TMH-B).
Fig. 8. Stress intensity factor versus strip-width to crack-length ratio
(TMB).
Fig. 9. Stress intensity factor versus magnetic field (h=a ¼ 2:0).
Y. Shindo et al. / Acta Materialia 52 (2004) 4677–4684 4683
Fig. 6 gives a comparison of the theoretical results of
kh1=r0a1=2 versus h=a with experimental data for TMC-
V. The theoretical results agree very well with the ex-
perimental data. Figs. 7 and 8 show the corresponding
results for TMH-B and TMB. Experimental measure-ments verify the predictions of a theoretical model. The
calculated kh1=r0a1=2 of TMC-V, TMH-B and TMB for
h=a ¼ 2 under various values of bc � 102 are compared
with the measured data in Fig. 9. A large value of bctends to increase the stress intensity factor depending on
the material.
6. Conclusions
The linear magneto-elastic problem for a soft ferro-
magnetic strip with a central crack has been analyzed
theoretically and the effect of the magnetic fields on the
stress intensity factor and energy density has been
summarized in a drawing. An experimental study has
also been conducted in which strain fields were used tocharacterize the stress intensity factor of a central crack
in nickel–iron soft magnetic materials with plate speci-
men geometry under the magnetic field. Both the theo-
retical and experimental results confirm the fact that the
applied magnetic field tends to intensify the fracture
machanics parameters such as stress intensity factor.
The excellent agreement between theoretical calculations
and measurements of stress intensity factor establishesthe validity of the linear theory for magneto-elastic
interactions.
Appendix A
fiðaÞ ði ¼ 1; 2; 3; 4Þ in Eqs. (31)–(34) are given by:
f1ðaÞ ¼R10
F1ðs; aÞ vB0
l0lrsAðsÞ � vB0
l0lr
2ðjþ2lÞjþl � sy
n oBðsÞ
h� lr �
2v2B20
l0l2r ðjþlÞ
n osaðsÞ
ids;
f2ðaÞ ¼R10
F2ðs; aÞsaðsÞds;
f3ðaÞ ¼R10
F1ðs; aÞ sAðsÞ � jþ2ljþl � sy
� �BðsÞ
hþ vB0ðjþ3lÞ
2llrðjþlÞ saðsÞids;
f4ðaÞ ¼R10
F2ðs; aÞ sAðsÞ � 2jþ3ljþl � sy
� �BðsÞ
hþ vB0ðjþ2lÞ
llrðjþlÞ saðsÞids;
9>>>>>>>>>>>>>>>>>>=>>>>>>>>>>>>>>>>>>;
ðA:1Þ
where
F1ðs; aÞ ¼ 2p
as2þa2 sinðshÞ;
F2ðs; aÞ ¼ 2p
ss2þa2 cosðshÞ:
�ðA:2Þ
4684 Y. Shindo et al. / Acta Materialia 52 (2004) 4677–4684
Appendix B
The functions EiðahÞ ði ¼ 1; 2Þ in Eq. (35) are given
by:
E1ðahÞ ¼ vB0
2llr
b5ðahÞ coshðahÞþðv�1Þb2ðahÞb1ðahÞ
f1ðaÞ � f2ðaÞf g� v2B2
0
2ll0l2r
b5ðahÞ coshðahÞþðv�1Þb2ðahÞb1ðahÞ
b4ðahÞb2ðahÞ
nþ b5ðahÞ coshðahÞ�b2ðahÞ
b2ðahÞ sinhðahÞ
of3ðaÞ
� v2B20ðjþ4lÞ
2ll0l2r ðjþlÞ
b5ðahÞ coshðahÞþðv�1Þb2ðahÞb1ðahÞ
n� sinh2ðahÞ
b2ðahÞ� b3ðahÞ
b2ðahÞ
of4ðaÞ;
E2ðahÞ ¼ � vB0
2llr
sinhðahÞ coshðahÞb1ðahÞ
f1ðaÞ � f2ðaÞf gþ 1
b1ðahÞb2ðahÞv2B2
0
2ll0l2rb4ðahÞ sinhðahÞ
nþ b1ðahÞ
ocoshðahÞf3ðahÞ
þ 1b1ðahÞb2ðahÞ
v2B20ðjþ4lÞ
2ll0l2r ðjþlÞ sinh
2ðahÞ coshðahÞn
� b1ðahÞosinhðahÞf4ðahÞ;
9>>>>>>>>>>>>>>>>>>>>>>>>>=>>>>>>>>>>>>>>>>>>>>>>>>>;ðB:1Þ
where
b1ðahÞ ¼ ah lr þv2B2
0ðj�lÞ
2ll0l2r ðjþlÞ
n ohsinhðahÞ þ coshðahÞ�
þ lr þv2B2
0ð2jþlÞ
2ll0l2r ðjþlÞ
n osinhðahÞ
hþ coshðahÞ� sinhðahÞ coshðahÞ;
b2ðahÞ ¼ ahþ sinhðahÞ coshðahÞ;b3ðahÞ ¼ ah coshðahÞ � l
jþl sinhðahÞ;b4ðahÞ ¼ ah� l
jþl sinhðahÞ coshðahÞ;b5ðahÞ ¼ ah coshðahÞ � 2l
jþl sinhðahÞ:
9>>>>>>>>>>>>>>>>=>>>>>>>>>>>>>>>>;ðB:2Þ
Appendix C
The functions LiðahÞ ði ¼ 1; 2Þ in Eq. (39) are given
by:
L1ðahÞ ¼ � b5ðahÞ coshðahÞþðv�1Þb2ðahÞ2b1ðahÞ
vbclr
� �2
c1
� b5ðahÞ coshðahÞþðv�1Þb2ðahÞ2b1ðahÞ
b4ðahÞb2ðahÞ
vbclr
� �2�
þ b5ðahÞ coshðahÞ�b2ðahÞb2ðahÞ sinhðahÞ
�c2
� b5ðahÞ coshðahÞþðv�1Þb2ðahÞ2b1ðahÞ
ðjþ4lÞ sinh2ðahÞðjþlÞb2ðahÞ
vbclr
� �2�
� b3ðahÞb2ðahÞ
�c3;
L2ðahÞ ¼ sinhðahÞ coshðahÞ2b1ðahÞ
vbclr
� �2
c1
þ 12b1ðahÞb2ðahÞ
b4ðahÞ sinhðahÞ vbclr
� �2�
þ 2b1ðahÞ�coshðahÞc2
þ 12b1ðahÞb2ðahÞ
ðjþ4lÞ sinh2ðahÞ coshðahÞjþl
vbclr
� �2�
� 2b1ðahÞ�sinhðahÞc3;
9>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>=>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>;
ðC:1Þwhere
c1 ¼ lr þ vbclr
� �2
;
c2 ¼ 12
vbclr
� �2
;
c3 ¼ 1jþ2l ðjþ lÞ þ l
2
vbclr
� �2� �
:
9>>>>>=>>>>>;
ðC:2Þ
References
[1] Shindo Y. Trans ASME J Appl Mech 1977;44:47.
[2] Shindo Y. Trans ASME J Appl Mech 1978;45:291.
[3] Pao YH, Yeh CS. Int J Eng Sci 1973;11:415.
[4] Shindo Y. Theor Appl Mech 1981;30:203.
[5] Yoshimura K, Shindo Y, Horiguchi K, Narita F. Fatigue Fract
Eng Mater Struct 2004;27:213.
[6] Dally JW, Sanford RJ. Exper Mech 1987;27:381.