In the present paper a simplified version of the transfer matrix method has been developed
for modal analysis of multiple cracked stepped beam based on an explicit expression of mode
shape of multiple cracked uniform beam element. The simplification consists of that the beam
state needs to be transferred only through steps of beam but not over the cracks as done in the
earlier publications.
An experimental modal analysis of cracked multistep beam has been carried out and
comparison of computed and measured natural frequencies demonstrated a good agreement of
the theory with experiment.
Using the simplified TMM it was found that likely to the uniform beam there exist on beam
positions crack appeared at which does not change a certain natural frequency. Such critical
points on beam are called herein frequency nodes and it was shown that step-down shifts the
nodes to the beam middle and step-up pulls them to the beam ends.
Finally, the performed modal analysis shows significant influence of steps on the natural
frequency sensitivity to cracks and this is a useful indication for crack detection in stepped beam
by measurement of natural frequencies
14 trang |
Chia sẻ: honghp95 | Lượt xem: 651 | Lượt tải: 0
Bạn đang xem nội dung tài liệu The transfer matrix method for modal analysis of cracked multistep beam - Vu Thi An Ninh, để tải tài liệu về máy bạn click vào nút DOWNLOAD ở trên
Vietnam Journal of Science and Technology 55 (5) (2017) 598-611
DOI: 10.15625/2525-2518/55/5/9140
THE TRANSFER MATRIX METHOD FOR MODAL ANALYSIS
OF CRACKED MULTISTEP BEAM
Vu Thi An Ninh1, Luu Quynh Huong2, Tran Thanh Hai3, Nguyen Tien Khiem3, *
1University of Transport and Communications, 3 Cau Giay, Dong Da, Hanoi
2Thuy Loi University, Chua Boc, Dong Da, Hanoi
3Graduate University of Science and Technology, VAST, 18 Hoang Quoc Viet, Cau Giay, Hanoi
*Email: ntkhiem@imech.vast.vn; khiemvch@gmail.com
Received: 9 January 2017; Accepted for publication: 28 May 2017
ABSTRACT
The present study addresses the modal analysis of multistep beam with arbitrary number of
cracks by using the transfer matrix method and modal testing technique. First, there is conducted
general solution of free vibration problem for uniform beam element with arbitrary number of
cracks that allows one to simplify the transfer matrix for cracked multistep beam. The
transferring beam state needs to undertake only at the steps of beam but not through crack
positions. Such simplified the transfer matrix method makes straightforward to investigate effect
of cracks mutually with cross-section step in beam on natural frequencies. It is revealed that
step-down and step-up in beam could modify notably sensitivity of natural frequencies to crack
so that the analysis provides useful indication for crack detection in multistep beam. The
proposed theory was validated by an experimental case study.
Keywords: stepped beam; cracked beam; modal analysis; transfer matrix method.
1. INTRODUCTION
Stepped beam structures have found widespread application in engineering fields such as
bridges, rotating machines, robotics and aerospace structures. In the engineering application,
vibration of the structures is the problem of a great importance and it is studied in the enormous
literature. Sato [1] studied an interesting problem that proposed to calculate natural frequency of
beam with a groove in dependence on size of the groove. Using a model of stepped beam and
Transfer Matrix Method (TMM) combined with Finite Element Method (FEM) the author
demonstrated that (a) fundamental frequency of the structure increases with growing thickness
and reducing length of the mid-step; (b) the mid-step could be modeled by a beam element,
therefore, the TMM is reliably applicable for the stepped beam if ratio of its length to the beam
thickness (r = L2/h) is equal or greater than 4.0. Comparing with experimental results the author
concluded that error of the TMM may be up to 20 % if the ratio is less than 0.2. Latter, Jang and
Bert [2, 3] used the conventional technique for calculating natural frequencies of two-step beam
and shown that natural frequencies of the structure are dependent not only on the change in
The transfer matrix method for modal analysis of cracked multistep beam
599
cross-section but also on the beam boundary conditions. Namely, stepping up (increasing height)
loads to increasing natural frequencies for any boundary conditions except the clamped ends
beam and stepping down (decreasing height) reduces the frequencies except the cantilevered
beam. The findings are important to show the dynamic property of stepped beam and which
method could be useful for vibration analysis of the beam. Other methods such as Adomian
Decomposition Method (ADM) and Differential Quadrature Element Method (DEM) have been
developed in [4] and [5], respectively, for free vibration analysis of multi-step beams. Cunha and
Junior [6] investigated effect of elastic boundary supports on natural frequencies and mode
shapes of multiple stepped beam. Kukla and Zamojska [7] studied effect of axial force on natural
frequencies and longitudinal or torsional vibrations of stepped one-dimensional structures such
as bars or shafts were studied in [8] by using Distributed Transfer Function Method (DTFM).
Jaworski and Dowell [9] have compared different theoretical methods and beam theories used
for free vibration analysis of multistep cantilever beam with experimental results. It was shown
by the authors that there is disaggreement between theoretical and experimental results.
Wattanasakulpong and Charoensuk [10] studied one-step beam made of functionally graded
material.
Vibration of stepped beam structures with cracks have been also intensively examined due
to that cracks are potential to reduce the serviceability of a structure and in consequence may
lead to a serious accidence if it is not early detected. To detect cracks in a structure its vibration
analysis is crucially important. Nandwana and Maiti [11] have established frequency equation of
an n-step Euler-Bernoulli beam with single crack in a form of 4(n+1) order determinant and used
for crack detection by natural frequencies. Using TMM, Tsai and Wang [12] obtained frequency
equation for cracked multistep Timoshenko beam in much simplified form of 4x4-dimention
determinant that simplifies significantly computation of the beam’s natural frequencies.
Maghsoodi et al. [13] have obtained an explicit expression of natural frequencies through the
crack magnitudes for multistep Euler-Bernoulli beam that provide a system of linear algebraic
equations for crack detection from natural frequencies. Li [14] was able to conduct a recurrent
relationship between vibration modes of adjacent steps that is straightforward to obtain an
explicit expression of frequency equation for multiple cracked and stepped beam. The TMM is
completely developed and used for solving both the forward and inverse problems for multistep
Euler-Bernoulli beam with arbitrary number of cracks by Attar in [15]. However, in the latter
publication the transfer matrix is very complicated because it should be assembled not only at
the steps of beam but also over the crack sections.
This paper presents the TMM developed for modal analysis of cracked multistep beam
based on an explicit expression for mode shape of multiple cracked uniform beam element. This
enables to much simplify the transfer matrix of multiple cracked multistep beam compared to
that was developed in [15] and it is validated by an experimental study.
2. GENERAL SHAPE FUNCTION FOR MULTIPLE CRACKED BEAM ELEMENT
Consider a uniform beam element of length L; material density (ρ); elasticity (E) and shear
(G) modulus; section area hbA ×= and moment of inertia 12/3bhI = . Assume furthermore
that the beam is cracked at the positions Lee n <<<< ...0 1 and the cracks of depth naa ,...,1
are modeled by equivalent springs of stiffness nKK ,...,1 . The springs stiffness is calculated
from the crack depths using formulas given in Appendix. For the beam element, free vibration is
governed by equation
Vu Thi An Ninh, Luu Quynh Huong, Tran Thanh Hai, Nguyen Tien Khiem
600
4 4 4 2
4( ) / ( ) 0, / ,d x dx x A EIϕ λ ϕ λ ρ ω− = = (2.1)
that is solved under the conditions at the crack positions
)0()0( −=+ jj ee φφ ; )0()0();()0()0( −′′′=+′′′′′=−′′=+′′ jjjjj eeeee φφφφφ ;
( 0) ( 0) ( ) ,j j j je e eϕ ϕ γ ϕ′ ′ ′′+ = − + (2.2)
where njKEI jj ,...,2,1,/ ==γ called hereby magnitudes of the cracks. Introducing
Krylov’s functions
,2/)sin(sinh)(;2/)cos(cosh)(
;2/)sin(sinh)(;2/)cos(cosh)(
0403
0201
xxxLxxxL
xxxLxxxL
λλλλ
λλλλ
+=−=
−=+=
(2.3)
that are all continuous particular solutions of Eq. (2.1), we can prove that the functions
∑
=
=−+=
n
j
jkjkk kexKxLxL
1
0 ,4,3,2,1),()()( µ (2.4)
where
( )
( ) ( ) : 0; sinh sin( ) ( ) , 0,1,2,3,
20 : 0;
p
p S x x x xK x S x p
x
λ λ
λ
≥ +
= = =
<
(2.5)
p is derivative order and so-called damage parameters 4,3,2,1;,...,2,1, == knjkjµ are defined
as
∑
−
=
=−′′+′′=
1
1
0 ,4,3,2,1),()([
j
i
ijkijkjkj keeSeL µγµ (2.6)
are solutions of Eq. (2.1) satisfying also conditions (2.2).
Since functions (2.3) and function )(xS defined in (2.5) are continuous solutions of Eq.
(2.1), the functions (2.4) would satisfy also the equation except crack positions where they need
to satisfy conditions (2.2). Indeed, since
(0) (0) (0) 0; (0) 1,S S S S′′ ′′′ ′= = = = (2.7)
one has got
)0()()0()()0()0(
1
1
0
1
0 −=−+−=−++=+ ∑∑
−
==
jk
j
i
ijkijk
j
i
ijkijkjk eLeeSeLeeSeLeL µµ ;
)0()()0()()0()0(
1
1
0
1
0 −′′=−′′+−′′=−′′++′′=+′′ ∑∑
−
==
jk
j
i
ijkijk
j
i
ijkijkjk eLeeSeLeeSeLeL µµ ;
)0()()0()()0()0(
1
1
0
1
0 −′′′=−′′′+−′′′=−′′′++′′′=+′′′ ∑∑
−
==
jk
j
i
ijkijk
j
i
ijkijkjk eLeeSeLeeSeLeL µµ ;
).0()0(])()([)0(
)()0()()0()0(
1
1
0
1
1
0
1
0
−
′′+−′=−′′+′′+−′=
=+−′+−′=−′++′=+′
∑
∑∑
−
=
−
==
jkjjk
j
i
ijkijkjjk
kj
j
i
ijkijk
j
i
ijkijkjk
eLeLeeSeLeL
eeSeLeeSeLeL
γµγ
µµµ
Thus, general solution of Eq. (2.1) satisfying conditions (2.2) can be found in the form
)()()()()( 44332211 xLCxLCxLCxLCx +++=φ (2.8)
The transfer matrix method for modal analysis of cracked multistep beam
601
where functions 4,3,2,1),( =kxLk are determined in (2.4)-(2.6) and 4321 ,,, CCCC are
arbitrary constants would be found using boundary conditions for the beam. Using the
expression (2.8) one is able to calculate displacement, slope, moment and shear force
respectively as follows
44332211 )()()()()()( CxLCxLCxLCxLxxW +++=≡ φ ;
44332211 )()()()()()( CxLCxLCxLCxLxx ′+′+′+′=′≡Θ φ ;
1 1 2 2 3 3 4 4( ) ( ) ( ) ( ) ( ) ( ) ;M x EI x EIL x C EIL x C EIL x C EIL x Cϕ′′ ′′ ′′ ′′ ′′≡ = + + + (2.9)
443322111 )()()()()()( CxLEICxLEICxLEICxLEIxEIxQ ′′′+′′′+′′′+′′′=′′′≡ φ .
that can be rewritten in the matrix form
{ ( )} [ ( )] { },x x= ⋅V H C (2.10)
where vectors TxQxMxxWx )}(),(),(),({)}({ Θ=V , TCCCC },,,{}{ 4321=C and matrix
′′′′′′′′′′′′
′′′′′′′′
′′′′
=
)()()()(
)()()()(
)()()()(
)()()()(
)]([
4321
4321
4321
4321
xLEIxLEIxLEIxLEI
xLEIxLEIxLEIxLEI
xLxLxLxL
xLxLxLxL
xH . (2.11)
This representation (2.10) of beam state will be employed below to develop the transfer
matrix method for multiple cracked stepped beam.
3. TRANSFER MATRIX FOR STEPPED BEAM WITH MULTIPLE CRACKS
Let’s consider now a stepped beam composed of m uniform beam segments designated with
subscript j, j=1,2,,m. Namely, material and geometry constants of j-th beam segment are
denoted by jjjjj LhbE ,,,, ρ . Suppose that each of the beam segments contains a number ( jn )
of cracks represented by its position jji nie ,...,1, = and magnitude /ji j j jiE I Kγ = and a
crack of magnitude /j j j jE I Kα = occurs at joint of (j+1)-th and j-th segments.
Introduce state vector for j-th step as { }Tjjjjj xQxMxxWx )(),(),(),()( Θ=V defined in
(2.9) symbolized with subscript j. Therefore, continuity conditions at step joints are
1 1 1 1(0) ( ); (0) ( ) ( ); (0) ( ); (0) ( )j j j j j j j j j j j j j j jW W L L M L M M L Q Q Lα+ + + += Θ = Θ + = =
or
1(0) ( ) ( ), 1,2,..., ,j j j jL j mα+ = Γ =V V (3.1)
where / 1 /j j j j jE I Kα γ= = and
1 0 0 0
0 1 0( )
0 0 1 0
0 0 0 1
α
α
Γ =
. (3.2)
Vu Thi An Ninh, Luu Quynh Huong, Tran Thanh Hai, Nguyen Tien Khiem
602
On the other hand, using representation (2.10) the introduced state vector )(xjV could be now
rewritten as
( ) [ ( )] ,j j jx x=V H C (3.3)
where
′′′′′′′′′′′′
′′′′′′′′
′′′′
=
)()()()(
)()()()(
)()()()(
)()()()(
)]([
4321
4321
4321
4321
xLIExLIExLIExLIE
xLIExLIExLIExLIE
xLxLxLxL
xLxLxLxL
x
jjjjjjjjjjjj
jjjjjjjjjjjj
jjjj
jjjj
jH . (3.4)
The shape functions )(xL jk defined in (2.4) where beam constants are replaced by those with
subscript j. Specifically, the frequency parameter λ defined in (2.1) now is
4/12 )/( jjjjjj IEhbρωλ = .
It is easily to show that expression (3.2) yields
( ) ( ) (0)j j jL j=V T V ; 1( ) ( ) (0)j j jj L −=T H H . (3.5)
So, combining the relationship (3.5) with (3.1) for j =1, 2, , m one obtains finally
1( ) [ ]{ (0)};m mL =V T V (3.6)
1 1[ ] [ ( ) ( ) ( 1)... ( ) (1)]mm mα α−= Γ − ΓT T T T . (3.7)
Usually, conventional boundary conditions are expressed by
0 1 L{ (0)} 0; { ( )} 0,m mL= =B V B V (3.8)
where 0 , LB B are matrices of 2 4× dimension. For instance, if both ends of the beam are
clamped the boundary matrices get the form
0
0
1 0 0 0 1 0 0 0
;
0 1 0 0 1 0L Lα α
= =
B B .
with 0 , Lα α being magnitudes of possible cracks at the end clamps. Consequently,
1[ ( )] (0) 0,ω =B V (3.9)
where
=
TB
B
B
L
0)(ω . (3.10)
Eq. (3.5) would have nontrivial solution with respect to )0(1V under the condition
( ) det[ ( )]) 0,D ω ω≡ =B (3.11)
that is frequency equation desired for the stepped beam with cracks.
For instance, if the left end of beam is clamped and the other one is free, i. e. the beam is
cantilevered, the boundary conditions are
0)()()0()0( 11 ===Θ= mmmm LQLMW .
Therefore, the frequency equation (3.10) gets to be
The transfer matrix method for modal analysis of cracked multistep beam
603
33 34
33 44 43 34
43 44
( ) det 0,CF
T T
D T T T T
T T
ω
≡ = − =
(3.12)
where 4,3,2,1,, =kiTik are elements of the total transfer matrix [T] defined in (3.4). Similarly,
frequency equation of stepped FGM beam can be obtained as determinant of a 2x2 matrix for
other cases of boundary conditions such as simple supports or clamped ends. Namely, for simply
supported beam with
0)()()0()0( 11 ==== mmmm LMLWMW ,
frequency equation is
0det)( 14323412
3432
1412
=−=
≡ TTTT
TT
TT
DSS ω . (3.13)
For beam with clamped ends where
0)()()0()0( 11 =Θ==Θ= mmmm LLWW ,
one has got
0det)( 14232413
2423
1413
=−=
≡ TTTT
TT
TT
DCC ω . (3.14)
Solving the frequency equations gives rise natural frequencies ,...3,2,1, =kkω of the beam
that in turn allow one to find corresponding solution of Eq. (3.8) as 11 )0( VV kD= with an
arbitrary constant Dk and normalized solution 1V . Afterward, mode shape corresponding to
natural frequency kω is determined for every beam step as follows
kjjjjjjjjkjkjk CxLCxLCxLCxLDxWx ωω=+++== })()()()({)()( 44332211Φ ;
})]{1()...2()1([)]0([ 11 VTTTHC −−= − jjjj , mj ,...,2,1= . (3.15)
The arbitrary constant kD is determined by a chosen normalized condition, for example,
1)(max
),(
=xjkjx
Φ .
Thus, the free vibration problem for stepped beam with multiple cracks is completely
solved by the simplified transfer matrix method.
4. EXPERIMENTAL SETUP AND MODAL TESTING TECHNIQUE
In this section experimental modal analysis is accomplished for the stepped beam with
clamped ends as shown in Fig. 1. Geometry and material parameters of the beam models are
given in Table 1. 3210 ; 7855 / ; 0.3E MPa kg mρ ν= = = .
Crack is produced by saw cut with very small wide and different depth 0 %, 10 %, 20 %,
30 % and 40 % beam thickness at fixed positions on beam. Therefore, the saw-cut can be treated
as an approximate model of open transverse crack described in [16]. Three scenarios of cracked
beam are investigated: single crack at position 450 mm; double cracks at the positions 200 mm;
450 mm from the left end and triple cracks at positions 200 mm, 450 mm, 800 mm from the left
end. In the first scenario, single crack of various depth (10 – 40 %) is examined. The second
scenario is tested with various depth of crack at the first span and the crack at intermediate span
Vu Thi An Ninh, Luu Quynh Huong, Tran Thanh Hai, Nguyen Tien Khiem
604
has fixed depth of 40 %. The last crack scenario is carried out for the beam with three cracks of
equal depth 40 %.
Figure 1. Model of stepped beam used in experimentation.
Table 1. Geometrical dimensions and material properties of two-step (three span) beam.
Geometrical
parameters (mm)
Beam spans
1st span 2nd span 3rd span
Wide, b 20 20 20
Height, h 15.4 7.5 15.4
Length, L 315 400 315
Total length 1230
Material properties 3.0;/7855;210 3 === νρ mkgMPaE
Figure 2. Measurement system PULSE 360.
Figure 3. Experimental model with measurement points.
1 2 3 4 5 6 7
L1 L2 L3
L
The transfer matrix method for modal analysis of cracked multistep beam
605
The PULSE B&K360 system, Fig. 2a, is employed for gathering and processing measured
data. An impact hammer (Fig. 3) is used for generating an excitation at position A denoted by
)(ωX in the frequency domain and an accelerometer has been employed for measuring response
( )(ωY ) at the position B on the beam. Hence, the signal processor installed in the measurement
system provides Frequency Response Function (FRF) between the positions A, B calculated as
( )( ) ,( )
XY
AB
XX
SH
S
ω
ω
ω
= (4.1)
where )(),( ωω XYXX SS are auto- and cross correlation functions respectively of the signals X and
Y. Magnitude of the function (4.1) is shown for instance in Fig. 2b. Multiple measurement of
FRF is performed for varying positions of excitation and response and all the measured data
gathered should give rise the same modal parameters of testing structure.
In the theory of structural vibration, it was shown that the FRF (4.1) can be expressed in
term of natural vibration modes as
∑
= +−
=
n
AB i
BA
H
1 22 ][
)()()(
ℓ
ℓℓ
ℓℓ
ωζωω
φφ
ω . (4.2)
where
ℓℓ
ζω , are natural frequency and damping ratio respectively of mode ℓ and )(A
ℓ
φ -
normalized th−ℓ mode shape measured at position A. Moreover, analysis of the function (4.2) in
the frequency domain exhibited that in the case of small damping and sparse distribution of
natural frequencies the frequency response function reaches its local maximums at resonant
frequencies
...3,2,1,2/ˆ 22 =−= ℓ
ℓℓℓ
ζωω (4.3)
The damping ratio is represented by sharpness of the resonant peak that determined by
2 1
ˆ
ˆ( ) / 2 ,ζ ω ω ω= −
ℓ ℓ ℓ ℓ
(4.4)
where 12 , ℓℓ ωω are two frequencies in both sides of ℓωˆ defined by
2/)ˆ()()( 11 ℓℓℓ ωωω ABABAB HHH == . (4.5)
So that natural frequencies are determined from the measured data as
...3,2,1,2/ˆˆ 22 =+=∗ ℓ
ℓℓℓ
ζωω (4.6)
and results are given in Table 2 in comparison with the numerically computed ones.
5. RESULTS AND DISCUSSION
5.1. Theoretical validation
Note that in the case of uncracked beam the transfer matrix method proposed in section 3
leads to its classical version. This can be validated first by using the method for computing five
lowest eigenvalues ( 24 / , 1,2,3, 4,5k kA EI kλ ρ ω= = ) of a uniform beam with clamped ends.
Results of the computation compared to those obtained by the classical analytical method (see
Table 2) show that the transfer matrix method is really an exact method equivalent to the
Vu Thi An Ninh, Luu Quynh Huong, Tran Thanh Hai, Nguyen Tien Khiem
606
analytical one. Moreover, natural frequencies of an intact stepped beam calculated by the
classical transfer matrix method are given in the first row of Table 3. The results compared to
those obtained by FEM and measured demonstrate the fact that measured natural frequencies are
more closed (almost identical for three lower frequencies) to the analytical ones than FEM
results. This validates reliability of measured data.
Table 2. Eigenvalues of uniform beam calculated by the TMM compared to analytical method [17].
Eigenvalues λ1 λ2 λ3 λ4 λ5
TMM 4.7300 7.8532 10.9956 14.1372 17.2788
Analytical 4.7300 7.8532 10.9956 14.1371 17.2787
5.2. Experimental validation
Table 3. Comparison of calculated and measured natural frequencies of three-span stepped beam with
clamped ends.
Crack
scenarios
Natural frequencies (Hz)
1 2 3 4 5
Intact
beam
TMM 73.2781 144.5188 301.1640 529.0126 726.2999
Experiment 73.38 144.30 301.10 526.50 723.81
FEM 74.8296 146.6345 304.9121 530.7197 729.8934
Single
crack
TMM 72.4574 143.6455 294.4666 519.7713 721.3785
Experiment 72.31 143.70 294.40 517.56 714.88
Double
cracks
TMM 71.4113 143.0440 287.3652 491.1691 680.994
Experiment 71.63 143.70 290.90 493.06 689.81
Triple
cracks
TMM 70.9058 142.6157 285.9632 482.6887 671.1254
Experiment 71.06 142.90 287.50 480.06 674.69
Single crack at 450 mm; Double cracks at 200; 450 mm; Triple cracks at 200; 450; 800 mm
from the left end of beam and all the cracks are of equal depth 40 %.
Measurements of natural frequencies have been performed at 7 points (see Fig. 7) and
measured data are processed accordingly to that procedure presented in section 4.
In Table 3 there are depicted five lowest natural frequencies calculated and measured for
three crack scenarios described in the last row of the table. The results show that discrepancy
between calculated and measured natural frequencies is within 2 %. So the theoretical
development proposed above in the sections 2 and 3 is thus experimentally validated and it can
be surely used for analysis of crack effect on natural frequencies accomplished below.
5.3. Effect of crack position and depth
The transfer matrix method for modal analysis of cracked multistep beam
607
For analysis of crack effect on natural frequencies of stepped beam two types of the beam
are investigated. The beam of first type called down-stepped (B1S) is shown in Fig. 3 that was
examined also in the experiment. The other one has intermediate span of thickness greater than
that of the end spans and this type is called up-stepped beam (B2S). Both the types of stepped
beam investigated below are clamped at the ends and have the following configurations:
1 2 3 1 2 3 1 3 2
1 2 3 1 2 3 1 3 2
B1S: 1 ; 0.1 ; 0.15 ; 0.10 ;
B2S : 1 ; 0.1 ; 0.10 ; 0.15 .
L L L m b b b m h h m h m
L L L m b b b m h h m h m
= = = = = = = = =
= = = = = = = = =
First, ratios of three lowest natural frequencies (cracked to intact) are computed for the
beams with single crack of different depth from 10 % to 40 % and position running from the left
to the right ends through the steps. Results are shown in Fig. 4 where the frequency ratios of
beam B1 on the left and those of beam B2 on the right. Observing graphs presented in Fig. 4
allows one to make the following notations: (1) Likely to the uniform beam, there exist positions
on the stepped beam crack occurred at which does not change a certain natural frequency. Such
positions are called frequency node and they are given in Table 4 for uniform and stepped
beams. Obviously, natural frequency nodes are located symmetrically about the beam middle for
symmetric boundary conditions; (2) The frequency ratios undergo a jump when crack passing
beam steps (this means discontinuity of frequency variation due to crack at beam steps). Expanse
of the jumps is different for various modes and it is certainly dependent on height of the steps;
(3) Natural frequencies, as well known, are monotonically decreasing with growing crack depth.
Table 4. Frequency nodes of five lowest modes for uniform and stepped beams.
Mode
No Stepped beam B1 Uniform beam B0 Stepped beam B2
1 0.85-2.15 0.67-2.33 0.56-2.44
2 0.46-1.5 -2.54 0.4 - 1.5 - 2.6 0.38 - 1.5 - 2.62
3 0.3-1.16-1.84-2.7 0.28-1.07-1.93-2.72 0.27 - 0.94 -2.06 - 2.73
4 0.24-0.94-1.5-2.06-2.76 0.22-0.83-1.5-2.17-2.76 0.21- 0.76 - 1.5 - 2.24-
2.79
5 0.2-0.75-1.26-1.74-2.25-
2.8
0.18-0.68-1.23-1.77-2.32-
2.82
0.17-0.65-1.16-1.84-2.35-
2.83
Total beam length L = 3 m; Span length L1 = L2 = L3 = 1 m; Steps at 1 m and 2 m
5.1. Effect of beam steps and crack position
In this subsection, aimed to study effect of steps (abrupt change in beam height), two
uniform beams and two stepped beams (with changed both sizes of cross section) of the
following geometry are investigated in addition to the stepped beams considered in the previous
subsection.
1 2 3 1 2 3 1 2 3
1 2 3 1 2 3 1 2 3
BU1: 1 ; 0.1 ; 0.15 ;
BU2 : 1 ; 0.1 ; 0.10 ;
L L L m b b b m h h h m
L L L m b b b m h h h m
= = = = = = = = =
= = = = = = = = =
1 2 3 1 2 3 1 2 3
1 2 3 1 2 3 1 2 3
B3S: 1 ; 0.15; 0.1; 0.15 ; 0.15; 0.1; 0.15 ;
B4S : 1 ; 0.1; 0.15; 0.1 ; 0.1; 0.15; 0.10
L L L m b b b m h h h m
L L L m b b b m h h h m
= = = = = = = = =
= = = = = = = = =
Vu Thi An Ninh, Luu Quynh Huong, Tran Thanh Hai, Nguyen Tien Khiem
608
As the stepped beams B1S and B2S have uniform width, the beams B3S, B4S are stepped
in both sizes of cross section (width and height). The first three frequency ratios computed in
dependence on the crack position for the uniform and stepped beams are presented in Fig. 5.
0 0.5 1 1.5 2 2.5 3
0.94
0.95
0.96
0.97
0.98
0.99
1
Crack position
Fi
rs
t f
re
qu
en
cy
ra
tio
L1=L2=L3=1m, b1=b2=b3=0.1 h1=h3=0.15;h2=0.1
First Step Second Step
a /h = 10%
a /h = 20%
a /h = 30%
a /h = 40%
10%
20%
30%
40% 40%
30%
20%
10%
0 0.5 1 1.5 2 2.5 3
0.955
0.96
0.965
0.97
0.975
0.98
0.985
0.99
0.995
1
L1=L2=L3=1m,b1=b2=b3=0.1m;h1=h3=0.1m,h2=0.15m
Crack position
Fi
rs
t f
re
qu
en
cy
ra
tio
First Step
a /h = 30%
a /h = 20%
a /h = 10%
10%
20%
20%
10%
40% 40%
a /h = 40%
Second Step
30%30%
0 0.5 1 1.5 2 2.5 3
0.95
0.955
0.96
0.965
0.97
0.975
0.98
0.985
0.99
0.995
1
L1=L2=L3=1m, b1=b2=b3=0.1 h1=h3=0.15;h2=0.1
Se
co
n
d
fre
qu
en
cy
ra
tio
Crack position
10%
a /h = 40%
a /h = 30%
20%
30%
40%40%
First Step Second Step
a /h = 40%
a /h = 30%
40%
10%
20%
30%
40%
20% 20%
0 0.5 1 1.5 2 2.5 3
0.96
0.965
0.97
0.975
0.98
0.985
0.99
0.995
1
1.005
L1=L2=L3=1m,b1=b2=b3=0.1m;h1=h3=0.1m,h2=0.15m
Crack position
Se
co
n
d
fr
eq
u
en
cy
ra
tio
a /h = 40% a /h = 40%
40%40%
30%
10% 10%
20%
10%
20%20%
a /h = 30%
Second StepFirst Step
10%
30%
20%
a /h = 30%
0 0.5 1 1.5 2 2.5 3
0.955
0.96
0.965
0.97
0.975
0.98
0.985
0.99
0.995
1
L1=L2=L3=1m, b1=b2=b3=0.1 h1=h3=0.15;h2=0.1
Th
ird
fre
qu
en
cy
ra
tio
Crack position
Second StepFirst Step
a /h = 40%
a /h = 40%
30%
20%
10%
40%
40%
10%
20%
30%30%
20%
10%
0 0.5 1 1.5 2 2.5 3
0.955
0.96
0.965
0.97
0.975
0.98
0.985
0.99
0.995
1
L1=L2=L3=1m,b1=b2=b3=0.1m;h1=h3=0.1m,h2=0.15m
Th
ird
fre
qu
en
cy
ra
tio
Crack position
a /h = 30%
a /h = 40%a /h = 40%
30%
10%
20%
10%
30%
20%
10%
20%
First Step Second Step
40%
40%
Figure 4. Effect of position and depth of crack on natural frequencies of beams B1 (left) and B2 (right).
The transfer matrix method for modal analysis of cracked multistep beam
609
0 1 2 3
0.95
0.955
0.96
0.965
0.97
0.975
0.98
0.985
0.99
0.995
1
Crack position
Fi
rs
t f
re
qu
en
cy
ra
tio
B1S
B1SB1S
B2SB2S
B3S
Second StepFirst Step
BU2
B2S
BU1
B1S: b1=0.10;b2=0.10; b3=0.10m
h1=0.15;h2=0.10; h3=0.15m
B3S: b1=0.15;b2=0.10; b3=0.15m
h1=0.15;h2=0.10; h3=0.15m
B2S: b1=0.10;b2=0.10; b3=0.10m
h1=0.10;h2=0.15; h3=0.10m
B4S: b1=0.10;b2=0.15; b3=0.10m
h1=0.10;h2=0.15; h3=0.10m
B4SB4S
B4S
BU1: b1=0.10;b2=0.10; b3=0.10m
h1=0.15;h2=0.15; h3=0.15m
BU2: b1=0.10;b2=0.10; b3=0.10m
h1=0.10;h2=0.10; h3=0.10m
0 1 2 3
0.95
0.955
0.96
0.965
0.97
0.975
0.98
0.985
0.99
0.995
1
Se
co
n
d
fr
eq
u
en
cy
ra
tio
Crack position
B2S B2S
First Step Second Step
BU2: b1=0.10;b2=0.10; b3=0.10m
h1=0.10;h2=0.10; h3=0.10m
BU1: b1=0.10;b2=0.10; b3=0.10m
h1=0.15;h2=0.15; h3=0.15m
B2S: b1=0.10;b2=0.10; b3=0.10m
h1=0.10;h2=0.15; h3=0.10m
B1S: b1=0.10;b2=0.10; b3=0.10m
h1=0.15;h2=0.10; h3=0.15m
B3S: b1=0.15;b2=0.10; b3=0.15m
h1=0.15;h2=0.10; h3=0.15m
B1S B1S
B3S
B4S
B2SB2S
B4S
B3S
B3S B3S
B1SB1S
B4S: b1=0.10;b2=0.15; b3=0.10m
h1=0.10;h2=0.15; h3=0.10m
BU1
BU2BU2
BU1
B4S B4S
0 1 2 3
0.955
0.96
0.965
0.97
0.975
0.98
0.985
0.99
0.995
1
Th
ird
fr
eq
ue
nc
y
ra
tio
Crack position
BU1
First Step Second Step
B3S
B1S BU2
BU2
BU1
B2S
BU1
BU2
B2S
B4S
B2S
B3S B3S
BU1BU1
B4S
B4S B4S
B1S
B4S: b1=0.10;b2=0.15; b3=0.10m
h1=0.10;h2=0.15; h3=0.10m
BU2: b1=0.10;b2=0.10; b3=0.10m
h1=0.10;h2=0.10; h3=0.10m
B2S: b1=0.10;b2=0.10; b3=0.10m
h1=0.10;h2=0.15; h3=0.10m
B1S
BU1: b1=0.10;b2=0.10; b3=0.10m
h1=0.15;h2=0.15; h3=0.15m
B1S: b1=0.10;b2=0.10; b3=0.10m
h1=0.15;h2=0.10; h3=0.15m
B3S: b1=0.15;b2=0.10; b3=0.15m
h1=0.15;h2=0.10; h3=0.15m
Figure 5. Effect of beam thickness variaton (steps) and crack position on natural frequencies.
Vu Thi An Ninh, Luu Quynh Huong, Tran Thanh Hai, Nguyen Tien Khiem
610
Comparing the ratios computed for beam BU1 and BU2 allows one to find that increasing
thickness of uniform beam makes all its natural frequencies more sensitive to crack. This
highlights the well-known fact that more stiff beam is more sensitive to crack. However, the
increasing or decreasing thickness of only mid-span in stepped beam leads to diminish or
magnify the second frequency sensitivity to crack occurred at the span. So, steps in beam
thickness may increase or decrease natural frequency sensitivity to crack in dependence on
where crack is located and which frequency is considered. Graphs given in Fig.5 show also that
frequency nodes of stepped-down beam (B1) are thrusted to the beam middle and the nodes are
pulled away from the middle for stepped-up beam (B2). Nevertheless, the steps do not shift the
node located at the beam middle and uniformly increasing thickness of uniform beam does not
change the frequency nodes.
6. CONCLUSION
In the present paper a simplified version of the transfer matrix method has been developed
for modal analysis of multiple cracked stepped beam based on an explicit expression of mode
shape of multiple cracked uniform beam element. The simplification consists of that the beam
state needs to be transferred only through steps of beam but not over the cracks as done in the
earlier publications.
An experimental modal analysis of cracked multistep beam has been carried out and
comparison of computed and measured natural frequencies demonstrated a good agreement of
the theory with experiment.
Using the simplified TMM it was found that likely to the uniform beam there exist on beam
positions crack appeared at which does not change a certain natural frequency. Such critical
points on beam are called herein frequency nodes and it was shown that step-down shifts the
nodes to the beam middle and step-up pulls them to the beam ends.
Finally, the performed modal analysis shows significant influence of steps on the natural
frequency sensitivity to cracks and this is a useful indication for crack detection in stepped beam
by measurement of natural frequencies.
Acknowledgement. This work was completed with financial support from NAFOSTED of Vietnam under
Grant of number 107.01-2015.20.
APPENDIX
CALCULATION OF CRACK MAGNITUDE
The so-called crack magnitude introduced above is calculated as [16]
)/()1(6/ 020000 hahfKIE νpiγ −== ; (A.1)
).6.197556.401063.47
0351.332948.209736.95948.404533.16272.0()(
876
54322
0
zzz
zzzzzzzf
+−+
+−+−+−=
The transfer matrix method for modal analysis of cracked multistep beam
611
REFERENCES
1. Sato H. - Free vibration of beams with abrupt changes of cross-section. Journal of Sound
and Vibration 89 (1983) 59-64.
2. Jang S. K. and Bert C. W. - Free vibration of stepped beams: exact and numerical
solutions. Journal of Sound and Vibration 130 (1989) 342-346.
3. Jang S. K. and Bert C. W. - Free vibration of stepped beams: higher mode frequencies.
And effect of steps on frequency Journal of Sound and Vibration 132 (1989) 164-168.
4. Mao Q. - Free vibration analysis of multiple-stepped beams by using Adomian
decomposition method. Mechanical and Computer Modelling 54 (2012) 756-764.
5. Wang X. W.and Wang Y. L. - Free vibration analysis of multiple-stepped beams by the
differential quadrature element method. Applied Mathematics and Computation 219
(2013) 5802-5810.
6. Cunha-Vaz J. and Lima-Junior J. J. - Vibration analysis of Euler-Bernoulli beams in
multiple steps and different shapes of cross section. Journal of Vibration and Control
(2014) DOI: 10.1177/1077546314528366.
7. Kukla S. and Zamojska I. - Frequency analysis of axially loaded stepped beams by
Green’s function method. Journal of Sound and Vibration 300 (2007) 1341-1041.
8. Yang B. - Exact transient vibration of stepped bars, shafts and strings carrying lumped
masses. Journal of Sound and Vibration 329 (2010) 1191-1207.
9. Jaworski J. W. and Dowell E. H. - Free vibration of a cantilevered beam with multiple
steps: Comparison of several theoretical methods with experiment. Journal of Sound and
Vibration 312 (2008) 713-725.
10. Wattanasakulpong N. and Charoensuk J. - Vibration characteristics of stepped beams
made of FGM using differential transformation method. Meccanica 50 (2015) 1089-1101.
11. Tsai T. C. and Wang Y. Z. Vibration analysis and diagnosis of a cracked shaft. Journal of
Sound and Vibration 192(3) (1996) 607-620.
12. Nandwana B. P. and Maiti S. K. - Detection of the location and size of a crack in stepped
cantilever beams based on measurements of natural frequencies, Journal of Sound and
Vibration 203(3) (1997) 435-446.
13. Maghsoodi A., Ghadami A. and Mirdamadi H. R. - Multiple crack damage detection in
multi-step beams by a novel local flexibility-based damage index. Journal of Sound and
Vibration 332 (2013) 294-305.
14. Li Q. S. - Vibratory characteristics of multi-step beams with an arbitrary number of cracks
and concentrated masses. Applied Acoustics 62 (2001) 691-706.
15. Attar M. - A transfer matrix method for free vibration analysis and crack identification of
stepped beams with multiple edge cracks and different boundary conditions. International
Journal of Mechanical Sciences 57 (2012) 19-33.
16. Chondros T. G., Dimarogonas A. D. and Yao J. - Longitudinal vibration of a continous
cracked bar. Engineering Fracture Mechanics 61 (1998) 593-606.
17. Nguyen Tien Khiem. Fundamentals of Structural Dynamics. VNU Publishing House,
Hanoi, 2004 (in Vietnamese).
Các file đính kèm theo tài liệu này:
- 9140_39555_1_pb_1265_2061431.pdf