CHANGE IN MODAL PARAMETERS OF CRACKED

∗
SINGLE-EDGE NOTCHED PLAIN CONCRETE BEAMS
Luis A. Padrón1 , Juan J. Aznárez1 , Héctor Cifuentes2 ,
Fernando Medina2 , Orlando Maeso1

1
University Institute SIANI, Universidad de Las Palmas de G.C., Las Palmas de G.C., Spain
{lpadron,jjaznarez,omaeso}@siani.es

2
Universidad de Sevilla, Seville, Spain
{bulte,medinaencina}@us.es

September 17, 2015

Abstract
The aim of this work is looking into the possibility of capturing the change in the modal
properties (natural frequencies, modal shapes and modal damping ratio) of plain concrete
elements due to the presence of cracked areas by using a simple continuum damage zone
numerical model. To do so, and as a first step, single-edge plain concrete beams were
identified by experimental modal analysis with and without cracks in order to find out
how cracking affects the first three flexural modes of the element. Then, numerical modal
analysis was performed by finite elements in order to identify the material properties of
the intact specimens by model updating and also check the experimental results. Finally,
a boundary element – finite element coupled code was used to find a continuum model
for the cracked single-edge notched beam able to reproduce the experimental results in
terms of natural frequency shifts and modal damping ratios. It was found that, in order
to reproduce the experimental frequency response functions, an overdamped damage zone
is needed together with an equivalent crack length.

Keywords: Modal analysis, Concrete, Crack Models.

1 INTRODUCTION
There exist mainly two types of approaches to study the behaviour of cracks in concrete
structural elements: discrete (i.e. fictitious crack model) or smeared cracking models (i.e.
continuum damage models). The first group of models is generally used to predict crack
growth and study the behaviour in the immediate vicinity of the damage. However, although
being more appealing for being closer to the real situation in plain concrete structures, discrete
crack models imply the modelling of individual cracks, which makes it not viable for large real
structures with multiple cracks. On the contrary, and specially in cases where the structure
is analysed under dynamic loads that do not induce crack growth, and when the points
Draft of the paper published in JOURNAL OF VIBRATION ENGINEERING & TECHONOLOGIES,
∗

3(3), 253-265 (2015)

1
of interest are not close to the cracks themselves, linear continuum damage models are a
viable alternative if they are capable of reproducing the general influence of the crack on
the response of the overall structure. This work focuses in that second situation in which a
concrete structural element is already cracked and, not being viable the modelling of every
individual crack, it is being submitted to dynamic loads (seismic, for instance, but in any case
not large enough to cause crack growth). As already well known [1, 2], the presence of those
cracks will modify the structural dynamic response, characterized by its modal parameters
(natural frequencies, modal shapes and modal damping ratios), and the estimation of such
modification is one of the aims of appropriate crack modelling.
On the other hand, countless works on damage identification can be found in the litera-
ture [3], many of which deal with crack location and severity in beams [4, 5, 6, 7, 8, 9, 10, 11,
12, 13, 14]. In this case, modelling the beams as 1D elements and the cracks as rotational
springs is a common approach. The focus of the research presented in this work is not on
damage identification or 1D modelling, but on cracking modelling in 2D or 3D solids and its
influence on the dynamic properties of the system.
In short, the aim of this work is looking into the possibility of capturing the change in the
modal parameters by using a simple 2D or 3D continuum damage zone numerical model for
cracked concrete elements. To do so, and as a first step, single-edge plain concrete beams are
identified by experimental modal analysis with and without damage in order to find out how
cracking affects the first three flexural modes of the element. Then, numerical modal analysis
is performed by finite elements in order to identify the material properties of the intact
specimens by model updating and also check the experimental results. Finally, a boundary
element – finite element coupled code is used to find a model of the single-edge notched
beam able to reproduce the experimental results obtained from the damaged specimens in
terms of natural frequency shifts and modal damping ratios. One of the parameters of such a
simplified model is the length of the equivalent notch needed to study the damaged specimen
under the linear-elastic hypotheses. The adequate value of this equivalent notch is found to
be coincident with the effective crack length given by the two-parameter fracture model for
concrete of Jenq and Shah [15]. It is also shown that, besides such equivalent notch length, a
high damping ratio in the linking fractured zone is also to reproduce both natural frequency
shifts and modal damping ratios.

2 METHODOLOGY
2.1 Overview
This work contains two distinct parts: a) an experimental part in which the modal parameters
of plain-concrete intact and cracked single-edge notched bending specimens are estimated;
and b) a numerical part involving different numerical models to address the problem at hand.
The details are given in the following sections.

2.2 Specimen description

The test specimens are prismatic single-edge notched bending plain concrete elements of
120 × 60 × 570 mm (height, width, length), with a center notch 3 mm width and 60 mm deep
(i.e. 50% of the specimen’s height). One specimen is shown in Figures 1 and 2.
Three specimens were produced and tested with a 1/1.7/2.9/0.5 (CEM-I-52.5 cement,
sand, crushed rock, water) concrete. The resulting concrete had a density (after 35 days) of
ρ = 2418.8 kg/m3 . They were dynamically identified (see section 2.3) after approximately

2
Figure 1: Specimen during a three-point flexural test (left) and detail of cracks after test
(right).

35 days. Immediately after identifying the intact specimens, they were subjected to a three-
point bending single-edge notched fracture test in order to produce visible cracks with lengths
around 30 mm at a first stage (damage state 1), and around 35 mm at a second and final stage
(damage state 2), but without reaching total fracture of the element (see Figure 1). Note
that those crack lengths given before are only rough estimates of the visible cracks observed
at both sides of the specimen, and are not a measure of the fracture process zone, which has
not been assessed. The state of the damaged specimen could be understood to be close to
that of a beam with a plastic hinge.

2.3 Experimental modal analysis description

The first three natural frequencies, modal shapes and modal damping ratios of the intact and
damaged specimens with free-free boundary conditions need to be experimentally identified
by experimental modal analysis. The characteristics of the specimen recommend the use of
the impact hammer technique [16, 17] to perform the experimental modal analysis. Free-free
boundary conditions were used, implemented by hanging the specimen by rubber bands. Since
4 accelerometers were available for the tests, and in order to be able to obtain acceleration
records in 7 different positions on the top edge of the specimen, two different setups (see
Figure 2) were used in the tests (one accelerometer stays in the same place to work as
reference sensor). To be sure that all significant modes were properly excited, the impacts
were applied on two different points (A and B in Figure 2). 5 repetitions were recorded for
each one of the two setups and two impact points, and for each one of the three specimens
in their three states. That would make a total of 180 impact hammer tests, but one of the
specimens could not be tested in the second damage state (because of premature failure), so
160 impact test were finally performed.
The technical characteristics of all equipment is detailed in the following: a) Impact
hammer: Brüel & Kjaer 8206-001. Reference sensitivity: 11.42 mV/N. Full Scale Force
Range Compression: 445 N. Effective seismic mass with enhancement: 140 gr. Aluminum
tip; b) Accelerometers: Brüel & Kjaer 4508-002 and 4507-002. Reference sensitivity: 1 V/g.

3
Figure 2: Experimental setups used in the experimental modal analysis by impact hammer.
Details of setup 1 (bottom left) and 2 (bottom right).

Frequency: 0.3-8000 Hz. Residual noise level (rms): ±0.35 mg. Mass: 4.8 gr. Beewax is used
for mounting; c) DAQ system: NI PXIe-4496, 24-Bit, Sigma-Delta ADCs, on NI PXIe-1073;
and d) Acquisition parameters: sampling frequency: 25000 Hz. Pre-trigger at 0.05 N with
100 buffer points.
Figure 3 shows an example of impulse recorded by the impact hammer (left) together
with its power spectral density estimation (centre) and a recorded time-history response.
The center figure shows that the energy provided to the system can be considered sufficient
up to 5000 Hz, while modes above 9000 Hz are not excited at all. The left figure shows
that the response damps out completely after around 0.3 s. For this reason, the response is
recorded during 0.4 s (10000 samples) and no windowing is necessary.

2.4 Numerical analyses

The calibration of the Young’s modulus of the concrete was made by matching the natural
frequencies of a numerical model with the experimental natural frequencies. To do so, modal
analysis was performed by finite elements using ANSYS R
software [18].
Numerical frequency response functions (FRFs), on the other hand, were obtained nu-
merically by a 2D BEM-FEM code in the frequency domain. The concrete specimen has
been modeled as an isotropic viscoelastic domain with hysteretic damping model of the type
E = Re[E](1 + 2iξ), being ξ the hysteretic damping ratio, and analyzed with the boundary

4
 350 −10 8

300 −20 6

Power/frequency (dB/Hz)
250 4
−30
200 2
−40

a (m/s )
2
F (N)

150 0
−50
100 −2
−60
50 −4

0 −70 −6

−50 −80 −8
3.5 4 4.5 5 0 2 4 6 8 10 12 0 0.1 0.2 0.3 0.4
t (s) −3
x 10 Frequency (kHz) t (s)

Figure 3: Characterization of the impulse applied for identification. Impulse (left), power
spectral density estimation (centre), and recorded time-history response (right).

element method [19] using quadratic elements. The rubber bands are modeled as Euler-
Bernoulli beams using 2-node 6-degrees of freedom beam finite elements [20]. The coupling
between boundary elements and finite elements is made through plane perfectly rigid surfaces
defined by any number of boundary elements, allowing the establishment of equilibrium and
compatibility conditions between those boundary elements and the finite element structure
attached to them.

2.5 Effective crack length by LEFM assumptions.

Several adaptations of the linear elastic fracture mechanics (LEFM) assumptions have been
proposed which take into account the nonlinear fracture behavior in an approximate man-
ner [21]. These adaptations are usually called effective crack models and they take into
account the pre-peak nonlinear behavior of concrete and are able to give a good estimate of
the peak load. One of the most universal effective crack model is the two parameter model
from Jenq and Shah [15]. On the basis of this method, an equivalent elastic structure contain-
ing an effective crack whose length is larger than the length of the crack in the real structure
should be determined [22]. Then LEFM can be applied to study the fracture behavior of the
concrete structure considering the effective crack length. In this way, the damage produced
in the fracture process zone ahead the real crack in the concrete structure is considered. The
model of Jenq and Shah will be compared here with the results by the BEM-FEM model in
terms of modal parameters.

3 RESULTS
3.1 Results from experimental modal analysis.
This subsection summarizes some of the results obtained from the experimental modal anal-
yses performed on the specimens. First of all, and just in order to illustrate the level of
repeatability of the experiments, figure 4 presents the moduli of the frequency response func-
tions, relating the response at ai2 to impact point A at setup 1, for the five tests on the
intact (left) and most damaged (right) states of the first specimen. In the case of the intact
specimen, the identified FRFs are very smooth and virtually independent on the test. As

5
expected, in the case of the damaged specimen, a small level of non-linearity is observed
when the results are analysed in detail. For instance, additional side peaks around resonance
(mostly below the second mode and around the first mode) appear when damage exists,
which suggests a non-linear behaviour of the cracks in the form of breathing cracks [9]. On
the other hand, and only for the most damaged states, for changes of around 60% in the im-
pact level, natural frequencies vary up to a 0.4% and peak magnitudes up to a 30%. For the
intact specimens, these non-linearities are virtually nonexistent. It is also worth mentioning
that the identified parameters can be considered consistent across repetitions and also across
specimens, which suggests that the experimental results are robust. Therefore, and for the
sake of clarity, results corresponding only to individual tests on specimen one will be shown
in this section.

0 0
10 10

−1 −1
10 10
|H3B(w)|

|H4A(f)|

−2 −2
10 10

test 1 test 1
test 2 test 2
−3 −3
10 test 3 10 test 3
test 4 test 4
test 5 test 5
−4 −4
10 10
0 1000 2000 3000 4000 5000 6000 0 1000 2000 3000 4000 5000 6000
f (Hz) f (Hz)

Figure 4: Illustrations of levels of repeatability and non-linearity in the estimated frequency

response functions. Setup 1. Measured at 60 mm left from the center (ai2). Impact on point
A. Intact specimen (left) and damaged specimen (right).

Table 1 presents the experimentally identified natural frequencies for the intact and dam-
aged states. Figure 5 shows the normalized frequency response functions corresponding to
an impact on point A (i.e. 30 mm to the left of the center) and measuring in channel ai2
from setup 1 (i.e. at 60 mm to the right of the center, see Figure 2). Real and imaginary
parts are also shown. As expected, frequencies are lower when the element is damaged due
to the smaller stiffness of the damaged zone. It can also be seen that, while the fundamental
flexural frequency gets reduced by more than 50% from the intact to the first damage state,
the second and third frequencies are reduced by only 8 and 2% respectively, showing that,
in this specific configuration of the element, the damage affects mainly to the fundamental
mode. The difference between the two damage states is much smaller. It is also apparent
how the first mode is extremely damped in the damaged specimens, due to the fact that most
of the deformation corresponding to that fundamental mode occurs in the central damaged
part.
Figure 6 shows the three modal shapes experimentally identified for the different damage
states. It is worth mentioning here that the modes do not change significantly between
intact or damaged specimens in terms of shapes, although they obviously change in terms of
amplitude and frequency. The first mode shape is only shown for the intact state due to the

6
 1 0.6
Real part
0.4 Imaginary part
0.8

Real[FRF(f)], Imag[FRF(f)]
0.2
Normalized |FRF(f)|

0.6 0

−0.2
0.4 −0.4

−0.6
0.2
−0.8 Intact specimen
0 −1
0 1000 2000 3000 4000 5000 0 1000 2000 3000 4000 5000
f (Hz) f (Hz)

1 0.6
Real part
Imaginary part
0.8 0.4
Real[FRF(f)], Imag[FRF(f)]
Normalized |FRF(f)|

0.6 0.2

0.4 0

0.2 −0.2
∼ 30 mm crack
0 −0.4
0 1000 2000 3000 4000 5000 0 1000 2000 3000 4000 5000
f (Hz) f (Hz)

1 0.5
Real part
0.4 Imaginary part
0.8
Real[FRF(f)], Imag[FRF(f)]

0.3
Normalized |FRF(f)|

0.6 0.2

0.1
0.4 0

−0.1
0.2
−0.2 ∼ 35 mm crack
0 −0.3
0 1000 2000 3000 4000 5000 0 1000 2000 3000 4000 5000
f (Hz) f (Hz)

Figure 5: Frequency response functions obtained from experimental modal analysis. Setup
1. Measured at 60 mm left from the center (ai2). Impact on point A. Intact (top), damage
state 1 (centre), and damage state 2 (bottom) specimens.

7
 1

0.5

−0.5

−1
0 0.2 0.4 0.6 0.8 1

Intact. f2=2960 Hz
1.5 ~30 mm. f2=2702 Hz
~35 mm. f =2652 Hz
2
1

0.5

−0.5

−1

−1.5
0 0.2 0.4 0.6 0.8 1

Intact. f3=4817 Hz
1.5 ~30 mm. f =4718 Hz
3
~35 mm. f3=4698 Hz
1

0.5

−0.5

−1

−1.5
0 0.2 0.4 0.6 0.8 1
x/L

Figure 6: Modal shapes obtained from experimental modal analysis. First (top), second
(center) and third (bottom) flexural modes.

8
 Mode 1 Mode 2 Mode 3
Intact specimen 760 2960 4817
damaged specimen (damage state 1) ∼300 2702 4718
damaged specimen (damage state 2) ∼300 2652 4698

Table 1: Experimentally identified natural frequencies (Hz).

mode overdamping observed for the damaged specimens.

Modal damping ratios were also estimated from the experimental frequency response
functions, finding a value of ξ = 0.005 (0.5%) for all the three modes of the intact specimen,
which supports the use of a frequency-independent hysteretic damping model for the numer-
ical analysis. In the damaged specimen, damping ratios for the second and third modes are
slightly larger while the fundamental mode is completely overdamped.

3.2 Results from finite elements modal analysis.

This section presents results from numerical modal analyses performed in order to identify the
material Young’s modulus by finite–element updating. The geometry of the intact specimen
was discretized by 3-D 10-node tetrahedral structural solid elements. Convergence analyses
regarding natural frequencies and modal shapes were made in order to define a valid mesh
for the analysis. A mesh with 29819 nodes and 19249 elements was found to provide accurate
results for the present analysis. The density of the concrete had been previously measured
to be ρ = 2418.8 kg/m3 . Poisson’s ratio was assumed to be ν = 0.2. As only one parameter
(material Young’s modulus) needed to be identified, and an estimation was already available
from static tests on different specimens, a simple ‘trial-and-error’ approach was used for the
model updating. A Young’s modulus E = 3.5×1010 Pa, very close to the previously available
static value for the concrete used for the specimens, was found to match the experimental
flexural fundamental frequency. Table 3.2 shows the very good agreement found also for the
higher modes. Figure 7 shows the corresponding modal shapes, that agree very well with the
experimental ones presented in Figure 6, giving more information about the deformation of
the element, not only on the top surface, but as a whole.

Mode Numerical (Hz) Experimental (Hz) Difference (%)

1 760 760 0
2 2917 2960 1.4
3 4651 4817 3.4

Table 2: Comparison between numerical (FEM) and experimental natural frequencies for the
intact state.

3.3 Results in the frequency domain from BEM-FEM analysis.

Using the material parameters found in the previous section (ρ = 2418.8 kg/m3 , ν = 0.2, E =
3.5 × 1010 Pa, ξ = 0.005), the two-dimensional BEM-FEM formulation described above has
been used to find a simple damage model for capturing the change in the modal parameters

9
Figure 7: Mesh and first three flexural modal shapes obtained from finite element modal
analysis for the intact state.

of the specimen due to the cracks. As shown above, such a model should be able of capturing
not only the frequency shifts but also the extremely high damping suffered by the first mode
in relationship with the the other modes. The geometry of the intact specimen was discretized
by 3-D 10-node tetrahedral structural solid elements. Convergence analyses regarding natural
frequencies and modal shapes were made in order to define a valid mesh for the analysis. A
mesh with 1196 BEM nodes, 588 BEM 2D quadratic elements, 8 FEM nodes and 6 beam
FEM elements was found to provide accurate results for the present analysis.
Experimental results for the intact specimen are very closely reproduced using the original
geometry and properties. On the other hand, in the case of the damaged configurations, a
model in which a link zone, geometrically modified in order to consider an equivalent crack
length larger than the actual crack (see section 2.5 and shadowed area Ω2 in Figure 8), is
given the same density and Young’s modulus than the rest of the element but a much higher
damping ratio (ξ2 = 2.00) has been found to reproduce very closely the experimental results
in terms of both frequency shift and overdamping of the first mode. Equivalent notch lengths
of 90 mm and 94 mm are found to provide the best approximations for both modal frequencies
and dampings of both damage states.

Mode 1 Mode 2 Mode 3

Intact specimen (exp.) 760 2960 4817
Intact specimen (num.) 760 2910 4650
Damage state 1 (exp.) ∼300 2702 4718
90 mm notch & ξ2 = 2.0 (num.) ∼300 2535 4610
Damage state 2 (exp.) ∼300 2652 4698
94 mm notch & ξ2 = 2.0 (num.) ∼300 2473 4606
94 mm notch & ξ2 = 0.005 (num.) ∼300 2578 4646

Table 3: Numerical BEM-FEM natural frequencies vs experimental natural frequencies (Hz).

Table 3 summarizes these results in terms of frequencies, while Figure 9 shows the normal-

10
 60 mm 30 mm FEM beam
elements
accelerometer F(w)
A
BEM
intact elements
specimen
model notch lenght (60 mm)

A
damaged
specimen equivalent crack length
model

Figure 8: BEM-FEM models for intact (top) and damaged (bottom) specimens.

1 1
Experimental Experimental
BEM−FEM BEM−FEM
94 mm equivalent
0.8 0.8
Intact specimen notch. ξ2 = 0.005
Normalized |FRF(f)|

Normalized |FRF(f)|

0.6 0.6

0.4 0.4

0.2 0.2

0 0
0 1000 2000 3000 4000 5000 0 1000 2000 3000 4000 5000
f (Hz) f (Hz)
1 1
Experimental Experimental
BEM−FEM BEM−FEM
90 mm equivalent 94 mm equivalent
0.8 0.8
notch. ξ2 = 2.0 notch. ξ2 = 2.0
Normalized |FRF(f)|

Normalized |FRF(f)|

0.6 0.6

0.4 0.4

0.2 0.2

0 0
0 1000 2000 3000 4000 5000 0 1000 2000 3000 4000 5000
f (Hz) f (Hz)

Figure 9: Numerical normalized frequency response functions for different models by BEM-
FEM formulation. Measured at 60 mm left from the center (ai2). Impact on point A.
Comparison with experimental results from intact (top left), damage state 1 (top right and
bottom left) and damage state 2 (bottom right).

11
ized frequency response functions for most of the described cases. The comparison between
these numerical results against the experimental results discussed above, shows the good
agreement between this simple model and the experiments. The match is not perfect, but
the general trends are indeed captured. Moreover, last row of the table 3 and top right plot
of Figure 9 show that the largest equivalent notch length and no additional damping provides
the best estimation for the second and third natural frequencies. Such a model, however, is
not able to capture properly the overdamping of the first mode due to the damage, or the
relationship between the relative peak values of the different modes.

3.4 Equivalent elastic crack length according to Jenq and Shah

Critical effective crack length ae according to LEFM was obtained with P-CMOD curves in
order to compare it with the results of the previous sections and see if these equivalent linear
approximations could be used in the future for more complex cases. This was calculated
using the compliance method of Jenq and Shah [15]. According to these indications, after
obtaining the initial compliance Ci of the P-CMOD curves, the elastic modulus of concrete
is obtained as Ec = 6Sao V1 (α0 )/(Ci BD2 ), where V1 (α0 ) is a geometric function given by

0.66 4D
V1 (α0 ) = 0.8 − 1.7α0 + 2.4α02 + 0
+ (−0.04 − 0.58α0 + 1.47α02 − 2.04α03 ), (1)
(1 − α ) 2 S

B is the specimen width, D is the depth of the cross section of the specimen, S is the
specimen loading span, L is the specimen length, a0 is the initial relative notch depth and
α0 = (a0 + HO)/(D + HO) , in which HO is the clip gauge holder thickness (Figure 10).
In this case, HO was 1 mm and there was no accuracy loss if α0 ≈ α was considered (being
α = a0 /D the relative notch depth). The value obtained for V1 (α) was 2.972 and the elastic
modulus of concrete was determined.

Figure 10: Details of the CMOD instrumentation for the three-point bending tests

Now, the critical effective crack length (ae ) is calculated from Ec and the unloading
compliance Cu , which was measured at 95% of the maximum load and assuming the unloading
path return to the origin. The critical effective crack length is found when the relationship
Ec = 6Sae V1 (αe0 )/(Cu BD2 ) is satisfied. This way, the obtained unloading compliance from
three-point bending tests was 6.477×10−5 mm/N ±15% and the obtained value of the critical

12
crack was ae = 90 mm, which coincides exactly with the result of the previous sections once
damping has been taken into account. It must be noted that this effective crack length
considers, by means of a longer crack length, the stresses relaxation and the stiffness reduction
produced in the fracture process zone of quasi-brittle materials (like concrete) when linear
elastic models are used.

4 CONCLUSIONS
This work presents a simple continuum damage zone numerical model for cracked concrete
elements to capture the change in modal parameters due to damage. A very specific case has
been investigated: a single-edge notched plain concrete bending specimen, which leads to an
specific configuration close to plastic hinge when damaged.
Several specimens were experimentally identified with and without damage. Several nu-
merical analyses were performed and a model comprised by an equivalent crack length and
an overdamped zone is proposed. It is shown that such a model is able to capture the main
features of the problem, in terms of both modal frequency shifts and modal dampings. To do
so, the equivalent crack length proposed by Jenq and Shah [15] (which allows the static anal-
ysis of the damaged specimen under the linear-elastic hypotheses) can be used. In addition to
this, in order to be able to reproduce the dynamic response of the system, overdamping must
be assigned to the rest of the damaged zone. In any case, in order to be able to draw general
conclusions, future developments should involve different and more general configurations
and sizes, more damage levels and different concrete mixes.

Acknowledgements
This work was supported by Subdirección General de Proyectos de Investigación of the Min-
isterio de Economı́a y Competitividad (MINECO) of Spain through the research project
BIA2010-21399-C02-01/02 (coordinated between the Universidad de Las Palmas de Gran
Canaria and the Universidad de Sevilla) and co-financed by the European Fund of Regional
Development (FEDER). The authors would also like to thank Jacob Montesdeoca for his
help in the preparation of specimens and tests; and Rubén Darias, Jacob Rodrı́guez, Ariel
Santana and Ayoze Méndez for their help during the experiments. The three-point flexural
tests were performed in the Laboratorio de Construcción of the Escuela de Arquitectura of
the Universidad de Las Palmas de Gran Canaria. The specimens and auxiliary equipment
were made in the Laboratorio de Hormigones, Tierras y Asfaltos of the EIIC of the same
university.

References
[1] P. Cawley and R.D. Adams, Location of defects in structures from measurements of
natural frequencies. J Strain Anal Eng Des, 14(2), 49–57, 1979.

[2] P. Gudmundson, Eigenfrequency changes of structures due to cracks, notches or other

geometrical changes. Journal of the Mechanics and Physics of Solids, 30(5), 339–353,
1982.

13
 [3] Q.W. Yang, Review of vibration-based structural damage identification methods. Journal
of Vibration and Shock, 26(10), 86–91, 2007.

[4] P.F. Rizos, N. Aspragathos and A.D. Dimarogonas, Identification of crack location and
magnitude in a cantilever beam from the vibration modes, Journal of Sound and Vibra-
tion, 138(3), 381–8, 1990.

[5] A. Morassi, Crack-induced changes in eigenparameters of beam structures. Journal of

Engineering Mechanics, 119(9), 1798–803, 1993.

[6] Y. Narkis. Identification of crack location in vibration simply supported beams. Journal
of Sound and Vibration, 172(4), 549–558, 1994.

[7] F. Vestroni and D. Capecchi. Damage evaluation in cracked vibrating beams using ex-
perimental frequencies and finite element models. Journal of Vibration and Control, 2,
69–86, 1996.

[8] M. Chati, R. Rand and S.J. Mukherjee. Modal analysis of a cracked beam. Journal of
Sound and Vibration, 207, 249–70, 1997.

[9] S.M. Cheng, X.J. Wu, W. Wallace W and A.S. Swamidas. Vibrational response of a
beam with breathing crack. Journal of Sound and Vibration, 225(1), 201–8, 1999.

[10] P.N. Saavedra and L.A. Cuitio. Crack detection and vibration behavior of cracked beams.
Computers and Structures, 79(16), 1451–6, 2001.

[11] J. Fernández-Sáez and C. Navarro, Fundamental frequency of cracked beams in bending

vibrations: An analytical approach. Journal of Sound and Vibration, 256(1), 17–31,
2002.

[12] A. Nandi and S. Neogy. Modelling of a beam with a breathing edge crack and some
observations for crack detection. Journal of Vibration and Control, 8(5), 673–93, 2002.

[13] A. Chatterjee. Structural damage assessment in a cantilever beam with a breathing

crack using higher order frequency response functions. Journal of Sound and Vibration,
329(16), 3325–34, 2009.

[14] F.B. Sayyad and B. Kumar. Theoretical and experimental study for identification of
crack in cantilever beam by measurement of natural frequencies. Journal of Vibration
and Control, 17(8), 1235–1240, 2011.

[15] Y.S. Jenq and S.P. Shah, A two parameter fracture model for concrete. Journal of
Engineering Mechanics, 111(4), 1227–1241, 1985.

[16] W.G. Halvorsen and D.L. Brown, Impulse technique for structural frequency response
testing. Sound and Vibration, November, 8–21, 1977.

[17] J. He and Z. Fu. Modal Analysis. Butterworth-Heinemann, 2011.

[18] ANSYS. Structural analysis guide – release 11.0, ANSYS Inc., 2007.

14
[19] J. Domı́nguez, Boundary elements in dynamics, Computational Mechanics publications,
Southampton, UK, 1993.

[20] O.C. Zienkiewicz, R.L. Taylor and J.Z. Zhu, The finite element method: its basis and
fundamentals, Butterworth-Heinemann, 6th ed. 2005.

[21] B. Karihaloo. Fracture Mechanics and Structural Concrete. Longman Scientific and Tech-
nical. USA, 1995.

[22] RILEM TC89-FMT 1991. Determination of fracture parameters (KICs and CTODc) of
plain concrete using three-point bend tests. Materials and Structures, 23(38), 457–460,
1991.

15