Damage Identification in Concrete using Impact Nonlinear Reverberation Spectroscopy
(2013) FMS820 20132Mathematical Statistics
 Abstract (Swedish)
 Nonlinear acoustic methods has shown increasing potential in the identication of
damage in brittle materials such as concrete. Commonly, these methods focus on
one or several resonance frequencies of the material and on the relative change in
resonance frequency and attenuation with strain amplitude which has been shown to
allow one to identify the extent of the material damage. However, current methods
are technically complicated and time consuming.
In this study, we use parametric methods to model the free vibration response
from concrete after an impact excitation. The technique is termed Impact Nonlinear
Reverberation Spectroscopy (INRS) and mainly focus on the strongest resonance
frequency in the vibration response. The... (More)  Nonlinear acoustic methods has shown increasing potential in the identication of
damage in brittle materials such as concrete. Commonly, these methods focus on
one or several resonance frequencies of the material and on the relative change in
resonance frequency and attenuation with strain amplitude which has been shown to
allow one to identify the extent of the material damage. However, current methods
are technically complicated and time consuming.
In this study, we use parametric methods to model the free vibration response
from concrete after an impact excitation. The technique is termed Impact Nonlinear
Reverberation Spectroscopy (INRS) and mainly focus on the strongest resonance
frequency in the vibration response. The experiment is easily performed and
the measured vibration response contains several resonance frequencies. The signal
is shown to be well modelled as a discrete sum of timevarying amplitude polynomial
phase signals. Using this model, we propose a novel estimation algorithm based on
the higher order ambiguity function in combination with a relaxationbased least
squares tting procedure. We show that the strongest resonance frequency, in this
case the fundamental symmetric longitudinal mode, may be well modelled with an
exponential decaying amplitude with a polynomial attenuation and phase, both of
order three. The frequency and attenuation evolution with strain amplitude for the
strongest mode are then analysed and associated with the degree of damage.
In agreement with earlier studies, our result on real measurements show a frequency
decrease as well as an increase of attenuation as a function of the strain
amplitude. Furthermore, it is conrmed that a steeper slope of the amplitude dependence
on both frequency and attenuation can be associated with a higher degree
of damage. In addition, an analysis of the frequency dependence versus amplitude
indicates that the slope is dependent on the initial excitation amplitude. This is
not consistent with earlier studies and is probably due to a larger strain amplitude
used in the measurements, for which eects from slow dynamics becomes visible. (Less)
Please use this url to cite or link to this publication:
http://lup.lub.lu.se/studentpapers/record/4192020
 author
 Dahlén, Unn
 supervisor

 Andreas Jakobsson ^{LU}
 organization
 course
 FMS820 20132
 year
 2013
 type
 H2  Master's Degree (Two Years)
 subject
 language
 English
 id
 4192020
 date added to LUP
 20131206 08:20:52
 date last changed
 20140304 15:31:21
@misc{4192020, abstract = {Nonlinear acoustic methods has shown increasing potential in the identication of damage in brittle materials such as concrete. Commonly, these methods focus on one or several resonance frequencies of the material and on the relative change in resonance frequency and attenuation with strain amplitude which has been shown to allow one to identify the extent of the material damage. However, current methods are technically complicated and time consuming. In this study, we use parametric methods to model the free vibration response from concrete after an impact excitation. The technique is termed Impact Nonlinear Reverberation Spectroscopy (INRS) and mainly focus on the strongest resonance frequency in the vibration response. The experiment is easily performed and the measured vibration response contains several resonance frequencies. The signal is shown to be well modelled as a discrete sum of timevarying amplitude polynomial phase signals. Using this model, we propose a novel estimation algorithm based on the higher order ambiguity function in combination with a relaxationbased least squares tting procedure. We show that the strongest resonance frequency, in this case the fundamental symmetric longitudinal mode, may be well modelled with an exponential decaying amplitude with a polynomial attenuation and phase, both of order three. The frequency and attenuation evolution with strain amplitude for the strongest mode are then analysed and associated with the degree of damage. In agreement with earlier studies, our result on real measurements show a frequency decrease as well as an increase of attenuation as a function of the strain amplitude. Furthermore, it is conrmed that a steeper slope of the amplitude dependence on both frequency and attenuation can be associated with a higher degree of damage. In addition, an analysis of the frequency dependence versus amplitude indicates that the slope is dependent on the initial excitation amplitude. This is not consistent with earlier studies and is probably due to a larger strain amplitude used in the measurements, for which eects from slow dynamics becomes visible.}, author = {Dahlén, Unn}, language = {eng}, note = {Student Paper}, title = {Damage Identification in Concrete using Impact Nonlinear Reverberation Spectroscopy}, year = {2013}, }