Damage

Identification
Dipendra Gautam
 Structural damage
• Reduction of the structural bearing capacity during service life.
I • Any deviation in the geometric or material properties of a structure, which has an
adverse effect on the current or future performance of the structure, such as
O
undesirable stresses, displacements, or vibrations of the structure.
E • Reduction of bearing capacity is usually caused by degradation of materials,
T
structural components, or connections due to environmental or loading effects.
• All structures continuously accumulate damage during their service life.
C
Accumulated damage may cause severe consequences.
• Typical types of damage in civil engineering structures include cracks, fatigue, steel
corrosion, concrete spalls, scour, and deterioration.
Material level Global performance
 Causes of damage
• Failure of the material: corrosion, fatigue, plasticity, and cracking.
I • Flaws, voids, cracks, and weak spots caused during manufacture
O • Loss of structural connections, i. e. loose bolts, and broken welds.
E
• Improper assembly or misfits during construction.

T
Cause Type
C
Damage parameters

Location Extent and severity

I
Direct Baseline
O
inspection comparison
E

T Change in structural
NDE
C
state

A priori Stiffness change, f,

I
Credit: Ettouney
O and
Alampalli 2012
E

C
 NDT Techniques
1. Acoustic emission (AE)
• Damage events in a material or structure create stress waves propagating throughout the structure.
These stress waves are called acoustic emission with a frequency range typically from 20 kHz to 1.0
MHz. The properties of these stress waves depend on the nature of structural damage and the properties
I
of the structure. By analyzing the measured acoustic emission signals, structural damage can be
O detected, and the state of the structure can be estimated.
• Acoustic emission is initiated when energy is released from a local source due to these processes within
E a material or structure. Thus, this technique makes it possible to detect the crack growth in real time.
There are several important properties of AE signals, including amplitude, frequency content and rate
T of decay. The most common sensor type used in monitoring stress waves in materials is based on a
surface mounted piezoelectric crystal. When high accuracy of damage evaluation is needed, the
C number of sensors must be increased, and subsequently the amount of data output to the signal
processing system is also increased.
• The AE method is very powerful in detecting any damage mode up to the microscale. However, since
acoustic emission energy can be extremely small, this can be a severe limitation, especially when used
in an in situ structural monitoring where the background noise can interfere with the AE signals. Also,
this method is less capable of damage characterization and further damage evaluation.
 NDT Techniques
2. Ultrasound
• As an ultrasonic wave propagates through the object, the wave changes its form through reflection,
refraction and scattering. In ultrasound scanning, a transmitter transfers ultrasound waves into the
material and its signal is collected by a receiver, once the transmitter has passed through the material.
I
The transmitter and receiver can be simply placed on opposite surfaces of the material. Then, the
O signals collected by the receiver can be processed and the damage in the test object can be detected.
• There are various ultrasonic testing methods available for practical applications, such as impact echo,
E pulse echo, sonic vibration, guided waves and laser-based ultrasound.
• Ultrasonic testing has proved its effectiveness in a variety of applications to civil engineering
T structures, such as condition assessment of bridge decks, concrete evaluation for locating voids and
discontinuities and properties estimation of structural geometries and materials.
C
 NDT Techniques
3. Guided (Lamb) waves
• Guided waves are defined as stress waves forced to follow a path defined by the material boundaries of
the structure. The critical elements of the guided wave technique for damage detection include
transducers, signal processing methodology and arrangement of the transducer network to scan the
I
structure. Guided waves are excited and received in a structure using transducers for structural damage
O assessment.
• GW have advantages:
E 1. ability to inspect large structures while retaining coating and insulation
2. ability to follow complex shaped structures
T 3. ability to inspect the entire cross-sectional area of a structure
4. excellent sensitivity to multiple defects with high precision of identification.
C
 NDT Techniques
4. Thermography
• Thermal imaging is a subsurface defect detection method, based on temperature differences measured
on the investigated surface during monitoring by using infrared sensors or cameras. Since heat flow is
associated with material properties, it is possible to detect material damage by monitoring how heat
I
flows.
O • Thermography has shown great promise in the field of NDT for civil engineering structures. These
methods can be used for assessing various defects in civil infrastructure, such as fatigue, reinforcement
E corrosion and seismic damage. Thermal imaging needs only low load magnitudes, and it can be utilized
to validate stress distributions predicted by finite element model at the initial stage and to detect the
T spread of damage during failure.
C
 NDT Techniques
5. Electromagnetic methods
• Electromagnetic methods use the interaction between the electromagnetic flux and waves and the flaws
in the test material. Typical electromagnetic methods in practice include eddy current, static magnetic
field, magnetic particle and microwave techniques. These techniques are able to detect surface or
I
subsurface flaws with size less than 1 mm. Among these methods, eddy current and microwave
O techniques can be used for in-situ automated testing, without requiring pretest and posttest
preparations. Static magnetic field techniques detect flaws by measuring magnetic field perturbations,
E while microwave techniques can detect interior damage by measuring electromagnetic wave reflection
to indicate damage states. All electromagnetic methods can detect fatigue cracks and earthquake
T damage in metals. Microwave techniques can detect even more types of damage, such as reinforcement
corrosion, scour and bridge security.
C • Electromagnetic methods can be used in many ways in structural damage detection. Most of these
methods can only be applied to conductive materials, such as metals, depending on the type of test, and
they are limited to detecting surface or near surface flaws. They are suitable for direct sensing and
require close proximity, thus their usefulness in remote sensing applications for SHM is limited.
 NDT Techniques
6. Capacitive methods
• Capacitive methods are based on electromagnetic techniques, and they were initially used for assessing
water content in civil engineering materials such as concrete, masonry and soils. This technique uses
two or more electrodes on the surface of materials, and then applies a voltage between them. This
I
system forms a capacitor, and the changes in capacitance indicate internal constituents, such as material
O properties and moisture content.

C
 NDT Techniques
7. Laser Doppler Vibrometer
• The laser Doppler vibrometer (LDV) is a noncontact velocity transducer, based on the analysis of the
Doppler effect on a laser beam emerging from a solid surface. In order to apply this method to damage
detection, an approach using modal properties can be employed, by assuming that damage will cause
I
detectable changes in the modal properties. These measurements can then be used to extract modal
O parameters such as natural frequencies and mode shapes. The obtained modal parameters give essential
information for structural damage assessment. Comparing these parameters allows the location and
E extent of damage in the structure to be determined by using vibration-based damage identification
methods.
T • The laser Doppler vibrometer is very promising, since the method has a high spatial resolution of
measurement. It is a noncontact method and thus is easy to implement in practice. For practical
C applications, the LDV based structural condition assessment should operate under ambient excitation.
This type of excitation is more attractive since it allows modal analysis to be performed under service
condition of the structure and does not require any artificial exciters.
 NDT Techniques
8. Global positioning system
• Global positioning system technology is an emerging tool that can measure directly both static and
dynamic responses. GPS is a satellite- based positioning system that allows users in the field to
determine their location without the need to transmit and hence identify their position. GPS technology
I
can provide relative displacements measured at rates of 20 Hz and even higher up to 100 Hz. The
O accuracy of dynamic displacement measurements using GPS is at a centimeter to millimeter level and
at a maximum distance from the reference GPS receivers to the structure receivers of up to 30 km.
E These provide a great opportunity to monitor the displacement or deflection behavior of large civil
engineering structures in real time under ambient loading conditions.
T • Advantages:
1. line of sight is not required
C 2. one base unit can serve several rovers
3. the data can be collected and downloaded directly to a web-based system for remote interrogation
4. the technique can be used where access is difficult.
 NDT Techniques
9. Global positioning system
• Global positioning system technology is an emerging tool that can measure directly both static and
dynamic responses. GPS is a satellite-based positioning system that allows users in the field to
determine their location without the need to transmit and hence identify their position. GPS technology
I
can provide relative displacements measured at rates of 20 Hz and even higher up to 100 Hz. The
O accuracy of dynamic displacement measurements using GPS is at a sub-centimeter to millimeter level
and at a maximum distance from the reference GPS receivers to the structure receivers of up to 30 km.
E These provide a great opportunity to monitor the displacement or deflection behavior of large civil
engineering structures in real time under ambient loading conditions.
T • GPS techniques have several major advantages, including:
❖ line of sight is not required
C ❖ one base unit can serve several rovers
❖ the data can be collected and downloaded directly to a web-based system for remote
❖ interrogation
❖ the technique can be used where access is difficult
 NDT Techniques
10. Visual inspection
• Visual inspection is the simplest and most straightforward NDT method. Visual inspection is based on
observing the condition of a structure visually and attempting to detect its current state, according to
personal experience and available guidelines. In order to reduce costs and time, visual inspection
I
should be an integral part of any NDT project. Visual inspection has many advantages, including being
O very simple and quick, able to apply remotely and suitable for local to global inspection. However, the
inspection is qualitative and may require special equipment for access.
E

C
 NDT vs. SHM
NDT SHM
Nondestructive testing provides a wide range SHM is an emerging method for real time health
of powerful techniques for detecting damage evolution of engineering structures using monitored data.
I and assessing conditions for civil engineering SHM can provide more information about current state
structures. NDT techniques have been in use estimate and future performance prediction of a structure,
O and in continuous development for a long so as to make appropriate decisions for effective
time. management of the structure. SHM methods consists of
E key components, such as sensing network, system
identification, damage diagnosis and damage prognosis.
T
Sensing technology of NDT techniques is well SHM methods often use various types of sensors, and
C established and very advanced. A single type their sensing network can be automated through advanced
of sensor is typically used in NDT techniques. data transmission systems such as wireless
Operations of NDT techniques are usually communications. SHM methods can cover large areas,
manual and relatively simple. NDT techniques and a priori knowledge of damage location is not needed,
are generally localized in nature, and they which is particularly suited for large civil engineering
generally require a priori knowledge of the structures.
location of damage in a structure.
 NDT SHM
NDT vs. SHM
System identification approaches are not very System identification approaches are an essential part of
useful for NDT techniques, since these most SHM methods, and they are well developed for
approaches identify global structural behavior evaluating structural and/or modal parameters from
rather than behavior in a local area. ambient measurements. System identification approaches
have been implemented to many SHM methods in civil
structures
I
NDT techniques can provide localized SHM methods can offer global identification of damage
O identification of damage sites, when the type and in a structure. The damage identification by SHM
region of damage is available before methods often requires more information about the
E identification. The global structural information geometry and material properties of the structure, which
and analytical model are not required for these is often not easy and straightforward. The identification
T techniques because of their localized nature. procedure of SHM methods may need extensive
NDT techniques often need relatively simple computational efforts. Reliable and efficient algorithms
C
computational algorithms and provide high for damage identification based on SHM are still in
reliability for the characteristics of the damage. development.
Due to the short time event of NDT, stochastic SHM methods can assess the performance of the
modelling of structural capacity structural system due to their global nature. SHM
and loading conditions is usually not possible methods can give information for stochastic modelling on
the basis of continuous monitored data, which is useful
for reliability analysis of the structural system.
 Signal Processing
Most structural damage detection methods are based on measuring and analyzing
time signals collected from sensors. Such time signals can be input signals, such as
earthquakes, wind pressures or blast pressures, or output signals, such as
I accelerations, strains or displacements. Signals can be broadly divided into two types
O according to signal property: stationary and nonstationary. Stationary signals do not
change their characteristics with time, while non-stationary signals change their
E characteristics. Most signals acting on civil infrastructure are non-stationary, such as
T earthquakes and wind. If signals are analyzed in the frequency domain by
transforming from the time domain, the damage identification process is often more
C effective. Therefore, signal processing methods, typically Fourier based transforms,
wavelet transforms and Hilbert–Huang transform, are required for structural damage
detection.
 Signal Processing
1. Fourier Transform
A time signal can be represented in the frequency domain by various Fourier transforms. When any time
signal x(t) over time t is available, the most popular forward and inverse continuous Fourier transforms
(FT) can be expressed, respectively, as
I

T
The discrete Fourier transform is obtained by representing the continuous Fourier transforms in terms of
C summations. The Fourier transforms have many advantages, such as being applicable to various problems,
well understood and explicit inverse transforms. However, their major disadvantage is that the Fourier
transform spectra do not reveal any information regarding the time dependency of frequency content of
the signal. Thus, the Fourier transforms are not applicable for non-stationary signals, such as wind and
earthquake signals, or nonlinear problems.
 Signal Processing
2. Wavelet transform
To address the issues in the Fourier transforms, wavelet transforms have been developed for effective signal
processing. The wavelet transforms are a tool that can cut data, functions or operators into different frequency
components with a resolution matched to its scale. The forward and inverse continuous wavelet transform
I
(WT) for a time signal x(t) are defined, respectively, as
O

C The scaling factor controls the amplitude of the wavelet function, and the translation factor controls the
location of the wavelet function on the time scale. As the scaling factor changes, the effective width of the
wavelet function changes. Unlike Fourier transforms, wavelet transforms can be designed to meet the
requirements of particular problems.
Wavelet analysis is very suited to analyzing nonstationary signal, so it can be used as a feasible method for
processing a signal to construct the intended feature index of structural damage. Wavelet analysis has various
applications in structural damage detection, including singular signal detection, signal-to-noise separation,
frequency band analysis and so on.
 Signal Processing
3. Hilbert-Huang Transform
The Hilbert–Huang transform (HHT) has become a popular method for signal processing. In contrast to other
common transforms, such as Fourier transforms, HHT is an empirical approach that can be applied to a data
set. This approach is based on the Hilbert transform, which is defined, for an arbitrary time signal x(t) with
I
the Cauchy principal value P, as
O

E HHT is the result of the Hilbert spectral analysis and the empirical mode decomposition; it uses empirical
mode decomposition to decompose a signal into so called intrinsic mode functions with a trend. Then HHT
T applies Hilbert spectral analysis to the intrinsic mode functions to obtain instantaneous frequency data. The
HHT keeps the characteristics of the varying frequency, since the signal is decomposed in the time domain,
C and the length of the intrinsic mode functions is the same as the original signal. This is a great advantage of
HHT since real world signals usually have multiple causes happening in different time intervals.
HHT is very useful for nonstationary, nonzero mean and nonlinear real signals. It makes an empirical mode
decomposition of the time signal into narrow band components with zero mean. These components do not
have a specific analytical representation, but each component can be associated with a physical meaning.
These components can be related to the mode shapes of dynamic structural systems and therefore the
existence of damage.
 Signal Processing example
%%Convert a Gaussian pulse from the time domain to the frequency domain.
%%Define signal parameters and a Gaussian pulse, X.
Fs = 100; % Sampling frequency
t = -0.5:1/Fs:0.5; % Time vector
I L = length(t); % Signal length
X = 1/(4*sqrt(2*pi*0.01))*(exp(-t.^2/(2*0.01)));
O %%Plot pulse in time domain
subplot(1,2,1);
E plot(t,X)
title('Gaussian Pulse in Time Domain')
T xlabel('Time (t)')
ylabel('X(t)')
C hold on
 Signal Processing example
%%Use fft function to convert the signal to the frequency domain, first identify a new input length that is the next power
of 2 from the original signal length. This will pad the signal X with trailing zeros in order to improve the performance of
fft.
n = 2^nextpow2(L);
I %%Convert Gaussian pulse to frequency domain
Y = fft(X,n);
O %%Define frequency domain and plot unique frequencies
f = Fs*(0:(n/2))/n;
E P = abs(Y/n).^2;
subplot(1,2,2);
T plot(f,P(1:n/2+1))
title('Gaussian Pulse in Frequency Domain')
C xlabel('Frequency (f)')
ylabel('|P(f)|^2')
hold off

ASSIGNMENT: Compare cosine waves in the time domain and the frequency domain.