Measurement 129 (2018) 68–90

Contents lists available at ScienceDirect

Measurement
journal homepage: www.elsevier.com/locate/measurement

A review of some advanced sensors used for health diagnosis of civil T

engineering structures
⁎
Swagato Das , Purnachandra Saha
School of Civil Engineering, KIIT, Deemed to be University, Bhubaneswar, Odisha, India

A R T I C LE I N FO A B S T R A C T

Keywords: The developments in structural health monitoring techniques have led to the invention of various sensors that
Non-contact sensor can be effective damage indicator. Due to environmental or electromagnetic effects and need for constant energy
Strain sensors source, the traditional sensors are unable to provide accurate and continuous measurements. In light of these
Dynamic monitoring events, new and improved sensors have been developed, along with wireless technology, to assist the monitoring
Accelerometers
process. With the need of detecting more than one damage parameters, multiplexed sensors have been the main
Corrosion sensors
Multiplexed sensor
interest of researchers. This paper deals with the different sensors used for determination of strain, acceleration
MEMS and corrosion. A brief comparative study has been performed and presented in the following review paper.
Fiber optic sensors Multiplexed Fiber optics sensor have proved quite effective for SHM and proved to be a good competitor with
Optimal sensor placement other sensors. Optimum Sensor Placement technique developed for low level damage diagnosis is Iterated
Improved Reduced System (IIRS) Method.

1. Introduction transmission, storage or management and diagnosis of the structural

health. These techniques, with integrated remote sensing, use of smart
Structural Health Monitoring (SHM) of civil structures, as the name materials and computer biased knowledge system, allow a civil en-
suggests means a proper evaluation of the civil structural condition. gineer to observe the performance of large infrastructures; tunnels,
SHM based techniques are Non-Destructive Damage Evaluation (NDE) bridges, offshore platforms, railways and at other places where it is not
or Non-Destructive Testing (ND/T) techniques which are able to iden- possible for onsite field test [3]. The traditional sensors used for the
tify damage in a structure at an early stage and evaluate the damage purpose of health diagnostics were the electronic strain sensors, ac-
extent so that the engineers can carry out the maintenance with the celerometers, displacement transducers, which provided the global and
help of this data. The main functions of SHM are: 1) damage detection local parameters such as frequency, mode shapes, and frequency. These
or validating the performance of large-scale complicated structures, 2) methods, however, posed a problem in measurement due to the en-
accurate identification of damage location, 3) check the severity of vironmental interference.
damage or evaluation of health status and predicting the service life of The modern NDE methods for monitoring the structural parameters
the structure, and 4) real-time data collection and diagnosis of struc- are eddy current [4], radioscopy [5], acoustic emission (AE) [6], Lamb
tural condition [1]. Previously, the most common monitoring process waves [7], comparative vacuum [8], methods based on electromagnetic
carried out to evaluate damage was only by visual examination. With impedance principle [9] and on fiber optics [10]. The methods invol-
the development of technology, various other methods have been in- ving eddy current and radioscopy make use of expensive and large
troduced to evaluate the structural health. Non-destructive monitoring transducers which have to be installed in large quantities for a struc-
techniques used previously were the X-rays, thermography, radio- ture. Therefore these methods proved to be uneconomical [11]. Com-
graphy and many others which required direct access to all the parts of pared to these methods, SHM carried out with the help of accel-
the structure. This condition is not possible to achieve in case of com- erometers, strain gage, piezoceramic material sensor, fiber optics,
plex structures. SHM had a breakthrough when it was integrated with corrosion sensors and acoustic emission sensors are more commonly
different modern technologies such as; magnetic, electrical, thermal, used. These sensors, being economical and light, can be installed at a
and photic and computer science technology [2]. The modern SHM larger scale [2,5]. A huge amount of data is being generated which has
technique includes sensing technology, data collecting or acquisition, to be arranged accurately which is a costly and challenging process

⁎
Corresponding author.
E-mail address: [email protected] (S. Das).

https://doi.org/10.1016/j.measurement.2018.07.008
Received 30 January 2018; Received in revised form 2 July 2018; Accepted 4 July 2018
Available online 05 July 2018
0263-2241/ © 2018 Elsevier Ltd. All rights reserved.
S. Das, P. Saha Measurement 129 (2018) 68–90

[12]. With the improvement in SHM technologies, sensor fusion tech- frequency domain [30]. This measured frequency is then processed to
nique has been introduced in the field. This technique enhances the obtain the change in strain and hence structural diagnostic is per-
decision making performance, thus reducing the uncertainty, by the use formed. Another major damage in the structure is due to rebar corro-
of multiple sensors [13]. The complex data set obtained from multiple sion which can be detected with the help of embedded sensors such as
sensors has to be merged such that no relevant data is skipped out. The piezoceramic lead zirconate transducer and galvanic couple [31]. In
final analysis of SHM system is done based on increased safety, reduced order to incorporate all the sensing properties, i.e. for damage results
maintenance, economic approach, reduced design margins and possibly from strain gauge, accelerometer and corrosion sensor, a new type of
zero false alarm rate [14]. A state of the art review has been performed sensor-fiber optic sensors have been introduced and are the most widely
in this paper for the different sensors in order to obtain a uniform base used sensor of the present time. These optics are light, easy to handle,
of comparison among different sensing techniques. occupy little amount of space and are not affected by electromagnetic
The researchers are mainly focussing their attention on the devel- interference, unlike the other sensors [32]. The main advantage of these
opment of innovative sensing technologies for health diagnostic ap- sensors is multiplexing [33]. These sensors can be used for measure-
proach. For observing the structural behaviour, an appropriate sensor ment of a number of structural parameters; strain, temperature, hu-
network is required for efficient signal processing and damage calcu- midity, acceleration, corrosion. This technology has been beneficial in
lation from the extracted data [15]. The most basic health monitoring reducing the number of different types of sensors employed for SHM
technique ever developed was based on the modal properties the [34]. Sensor placement is also an important aspect for efficient sensor
structures: Vibration based SHM [16]. These methods include the use of placement such that maximum number of structural data is obtained
dynamic measurements to calculate the local and global behaviours of with deployment of minimum number of sensors. Study of Optimal
the structure. Non contact damage detection technique has been de- Sensor Placement (OSP) is advantageous for determining the efficiency
veloped to study the growth of cracks in structure. The local behaviour of sensor network [35].
is measured in terms of change in strain which can be measured with This paper deals with the development of SHM techniques in the
the help of strain-based sensors: strain gauge [17]. The global beha- field of sensor used. Different types of sensors used for SHM till date has
viours are measured as displacement, frequency and mode shapes been reviewed according to their loading and a brief comparative study
which can be measured with the help of sensors such as velocimeters, have been performed citing the advancement of sensors and their
accelerometers and displacement transducers [18]. The data collected sensing technology. A brief OSP study has also been conducted for
from these sensors are used for the mathematical calculations of da- finding the efficient sensor network placement.
mage in these structures using frequency response method, Modal As-
surance criterion method, modal strain energy and modal flexibility 2. Sensors vision based SHM
method and mode shapes and curvature [16,18–20]. Out of these
methods, frequency response method, which gives a precise location of Vision-based monitoring is a measurement technology used for de-
damage in global coordinates, is unable to measure local damage [21]. tecting damage in structure since the beginning of health monitoring. In
Other parameters, such as modal flexibility and modal energy, which normal monitoring, the mode shapes are calculated from the sensor
are theoretically more sensitive, pick up errors due to measurement and data, which involves a great number of calculations [16]. With an acute
environment along with the damage signals [22]. Some of the notable measurement of the vision-based system, the error sources were care-
damage algorithms developed for SHM of bridges is based on load fully examined such as gray level interpolation [36], subset function
ratings, ambient truck load, traffic loads or Weigh-in-Motion (WIM) [37] and shape functions [38]. With recent development, non-contact
data [23–25]. However the all these algorithms are dependent on the machine-based vision technology was introduced which can measure
strain data for the damage algorithms to detect damage [26]. Strain as a the quasi-distributed displacement of structures. The techniques in-
local coordinate has proved to be more sensitive in case of local damage volved in vision-based displacement sensors are digital image correla-
detection. The traditional strain sensors or strain gage can detect da- tion (DIC) [39], pattern matching [40], RANSAC algorithm [41] and
mage precisely only if the damaged area is present under the influence others.
or radius of the gage. This sensor is unable to pick up signals if the Machine-based vision technology includes measurement of strain
damage occurs in the surrounding areas outside the gage influence. from digital image obtained from high resolution industrial charge-
Therefore, the strain gage cannot be employed for complex civil coupled device (CCD) camera using multi-point pattern matching al-
structures [17]. Also, the sensors are vulnerable to surrounding elec- gorithm [42,43]. The pre-designed patterns are compared with the
tromagnetic interference. Thereafter the study of damage with the help images captured and the count is made for a number of times the
of piezoelectric sensor or transducers became popular. The PZT used patterns match with the images. When this count is maximum, it in-
are practically immune to the interference of electromagnetism. dicates that the target position is identified as damaged [44]. This
Therefore the damage recorded using this sensor is more precise than process is sensitive to the variation of amplitude of geometric size. After
the strain gauge sensor. The electromechanical impedance (EMI) the pattern matching is completed, the change in the image is converted
method and Lamb wave method utilizes mainly the piezometric trans- into actual structural displacement in order to measure the damage.
ducers (PZT) for monitoring the structure. In the former method, the One of the factors that have to be taken into consideration while
piezoelectric sensor can detect changes or damage in the mechanical obtaining images is the illumination which plays an important part as
properties of the structure by electrical impedance recorded by the inferior illumination may lead to images captured not real to the ori-
sensor [27]. Whereas the latter method utilizes PZT to monitor Lamb ginal structure. The effect of bad weather, fog, may also be a hindrance
waves which can travel long distances with a little loss in amplitude to measurement. The vision-based measurement was tested against the
[28]. This method can be efficiently utilized if the number of PZT used magnetostrictive displacement sensor (MDS). The lab experiments for
is more [29]. The above sensors are mainly used to measure the damage these factors proved to be satisfactory. However, in case of field ex-
due to change in strain that occurs in the structure after a particular periments, a large angle of elevation for the digital camera has to be
force has acted on it, ex. earthquake force. Such type of damage is taken keeping in mind the sensitivity and accuracy of the measuring
normally referred to as the static damage in the structure. For struc- system. For a large angle of elevation, the longitudinal axis of the di-
tures, such as towers or bridges, continuous horizontal force, in form of gital camera may not be perpendicular to the object, hence causing a
wind load, is acting. For assessment of such structures, dynamic para- problem in measurement. The digital image processing can be used
meters have to be measured. This condition is achieved by the use of along vision-based technology for health monitoring and was compared
accelerometers. These sensors measure the acceleration of the structure with linear variable differential transformer (LVDT). The result was
continuously and diagnostic is done by converting this acceleration to found to be reliable and effective for multipoint structural displacement

69
S. Das, P. Saha Measurement 129 (2018) 68–90

Fig. 1. Schematic of strain gauge shunt on the target surface [55].

measurement [45]. 3. Sensors for strain measurement in SHM

Side by side, video camera based vibration measurement was also
developed for NDT. The video-based monitoring can be applied for The principal parameter used to detect damage in a structure is the
analysis of small motions, small subpixel displacement, and the accu- change in strain. The change in strain may occur due to various reasons;
racy is more as compared to camera-based measurement [46]. Com- change in temperature, dynamic forces acting on the structure, tension
puter-based programming can be used to convert displacement units and compression forces due to loading or unloading and others. The
obtained in pixels to mm. The process of converting video signals to change in strain is measured after the deformations have occurred due
displacement signals was proved to be effective and was validated using to the application of external force or load.
signals measured from laser vibrometer and accelerometer [47]. In A wide range of sensors has been available for the purpose of SHM.
order to increase the signal to noise ratio, the video signals may be But most of the sensors have a problem of measurement due to en-
downsampled. Fast Fourier Transform is implemented to convert the vironmental disturbances. One of the major environmental issues is the
displacement signal to the frequency domain. The frequency obtained is change in temperature; due to which the structure experiences change
compared for health monitoring. in strain. Therefore it is hard to distinguish the strain measurements
The limitations are similar to that of the high-resolution camera- occurring due to change in temperature or due to change in other
based monitoring, i.e., illumination, hindrance due to weather condi- parameters. In order to measure this strain change, different strain
tions. However for non contact based monitoring the measurement of sensors have been introduced. There are mainly three different strain
noise is greater. Also for vibration of small amplitude, the sensor might sensors; electrical [48], wireless [49] and optical [50]. The electrical
now be as effective as strain gauges or accelerometer. strain gauges are inexpensive and suitable for dynamic load conditions.
Electrical resistance strain gauge (ERS) and vibrating wire strain gauge
(VWSG) comes under this category. VWSG has the ability to measure

70
S. Das, P. Saha Measurement 129 (2018) 68–90

temperature as well as strain without any electromagnetic interference. metal films and semiconductors [59]. The comparison of the above
But due to the limited sampling rate, VWSG are not suitable for long- materials is provided in Table 1. It can be seen that thick film resistor is
term monitoring of live load [51]. The wireless sensors require no cable a good bridge between the other two sensors. It is economical, reliable
for data transmission but their monitoring time is limited due to battery and has a good gauge factor which is important from a sensitivity point
life [49]. The optical sensors used are Fiber Bragg Sensors (FBG) and of view.
Fiber-Optics sensor (FOS). These sensors are immune to noise and To describe the sensitivity of a strain sensor, the gauge factor plays
electromagnetic interferences. Also, they can be safely integrated with an important role Eq. (1) [60].
the structure for long-term monitoring [52]. Commercially three types
ΔR/ R 0
of strain sensors are available; discrete strain sensor or point sensors, GF =
ΔL/ L0 (1)
quasi-distributed strain sensor and highly distributed strain sensor [53].
Discrete strain sensors can measure only localised strain, quasi-dis- where ΔR , ΔL , R 0 and L0 are change in resistance, change in displace-
tributed sensors are series of discrete sensors connected together and ment, initial resistance and length of composite sensor respectively.
highly distributed strain sensors can measure the complete profile of The gauge factor is calculated using Quarter Bridge configuration
strain along the entire length. Quasi-distributed sensors are a con- which gives the relation between unbalanced output voltage and strain
tinuous series of discrete sensors connected together for strain mea- due to the load applied (Eq. (2)) [61].
surement. Fibre Bragg Grating sensors (FBG) connected in series is the
most widely used quasi-distributed strain sensor. The highly distributed GF = (4V0/ V ) ∗ (1/ ε ) (2)
strain sensors, distributed optical fibre (FO) cables, provide strain data
where GF is the gauge factor, V is bridge excitation voltage, V0 is un-
over the whole length of the optical fibres. Therefore this sensor can be
balance bridge voltage due to the load applied and ε is the strain of the
used for long-term monitoring of structures as it measures the strain for
specimen.
the entire profile of the structure whereas the former sensors only
Out of all the TFSs, Thick Film Ceramic Strain Gauge Sensor (TFCS)
provided the strain data for the region they were installed in [54].
has been recognized as an effective strain measuring device. Ceramic
has high elasticity, shock resistant, corrosion and abrasion resistant, has
3.1. Metallic thin film strain gauge an operating thermal range from −20 °C to 80 °C and can withstand
high voltages (2 kW) [61].
The metallic thin film strain gauges are electrical sensors perfect for Jabir and Gupta [61] conducted a comparative Four-Point Bending
detection of a change in structural parameters and have been used for a Test (4PBT) experiment (ASTM D6272-00 2001) with the help of FSG
long period of time. These sensors are economical, compact, and stable and TFCS to determine the efficiency of the sensors. Each TFCS was
and provide a stable output signal for data acquisition. But the sensors connected to a Wheatstone bridge which in-turn was connected to a
have faced a challenge during measurement of dynamic loads of large data acquisition system. This arrangement is followed so that both
scale [55]. Thus these sensors are not suitable to measure the dynamic tensions on the upper side and compression on the lower side can be
strain of off-shore structures which can experience strain more than recorded for the beam. The response for TFCS having a GF 7.2 was
5000με. The normal metallic thin foil strain gauges are unable to observed to be higher than FSG having GF 2 (Fig. 2).
measure dynamic strain more than 2000με for 106 alternating load But for practical applications, the TFCS have not been calibrated for
cycles [56]. readings due to different atmospheric effects, disturbance due to the
To counter all the above drawbacks, strain shunt has been designed application of glue and continuous structural deterioration as the G.F.
Xia and Quial [55] to incorporate the metallic thin film sensor gauge for changes.
measurement of large dynamic strain. The design of strain shunt in-
cludes installation of a strain gauge on an elastic shunt sheet which is 3.3. Carbon based graphene sensor
supported by two elastic columns as shown in Fig. 1. Due to deforma-
tion in structure, the supporting columns will get deformed, thus dis- With the advancement in technology, new lightweight sensors have
turbing the strain gauge. Therefore the strain can be measured by the been developed. These sensors are based on carbon nanomaterials;
gauge without undergoing large deformations. It was experimented and carbon black (CB) [62], graphene [63], carbon nanotube (CNT) [64],
observed that the 11.5% strain was overestimated which may be due to carbon nanofiber (CNF) [65] and carbon nanoribbons. Among these
not considering the height of the supporting columns which may un- graphene has been the widely used nanomaterials. It has a 2D hex-
dergo bending and tension/compression. agonal honeycomb structure which easily breaks under stress, thus
changing the electronic energy band indicating damage in the structure
3.2. Thick film ceramic strain sensors [66].
Zha et al. [60] performed experiments on the piezoelectric strain
Measurement of strains in structures has been widely performed sensor based on graphene/epoxy materials (f-GnP) for SHM. The be-
with the help of Foil Film Strain Gauge or Strain Gauges (FSG) [57]. haviour of the sensors was studied under successive tensile loading and
With inclusions of new technology, Thick Film Sensors (TFSs) have incremental cyclic loading. From the results obtained, Fig. 3, it can be
been used for strain measurement. The thick film materials used are concluded that these composite sensors are effective for damage de-
chosen such that they have high longevity, more robust, survive high tection. From the graph obtained for tensile loading Fig. 3a, it can be
mechanical stresses and extreme temperatures [58]. These sensors also seen that with increased stress, the piezoresistivity increases and the
measure strain with the help of piezoresistive effect. The piezo-resistive damage is indicated by the irreversible nature of curve. Fig. 3b shows
materials mainly used for strain gauge are thick film resistors, thin the damage measured by incremental cyclic loading. It can be observed

Table 1
Characteristics of piezoelectric material strain gauge [59].
Material Gauge Factor Temperature Coefficient of Resistance – ppm/0C Temperature Coefficient of Gauge Factor- ppm/0C Stability

Metal sheets and films 1–2 20 100 Excellent

Silicon single crystals 50–80 1000 −1500 Good
Thick Film resistor 2–35 50–200 −300 Very Good

71
S. Das, P. Saha Measurement 129 (2018) 68–90

factors ensure good relative sensitivity and mechanical properties thus

ensuring the use of the sensor in the field.

3.4. Cement based strain sensor (CBSS)

Neild et al. [67] conducted practical and theoretical experiments to

conclude that for sensor embedded in massive concrete structure and
sensor on the surface, the readings may vary. In order to extract the
exact value of strain, concrete filled steel tubular (CFST) columns have
been used for SHM of bridge piers, multi-storeyed buildings, etc [68].
The health diagnostic of these columns is difficult due to the sur-
rounding steel tube. The traditional ultrasonic sensors or transducers
can be used for the purpose, but the quantity required may be high. The
Cement-based strain sensor (CBSS) is a type of high strength smart
sensors used for health monitoring of Ultra-high strength concrete
Fig. 2. Comparison of Thick-film ceramic sensor to FSG for tension and com- (UHSC) columns. Due to the cementitious nature, these sensors are
pression [61]. bonded naturally with the parent structure [69]. This sensor detects
damage by the piezoresistive effect, change in volume electric re-
sistivity, with the help of a fractional change in resistance as in Eq. (3)
[70]:-
ΔR/ R 0 = (Rx −R 0)/ R 0 (3)

Rx = Vx R/ VR (4)
where R 0 is the initial resistance of CBSS and Rx is the resistance of CBSS
during loading, Vx is the voltage measured across the specimen during
SHM process, VR is the voltage across the known resistor and R is the
known resistance of the resistor.
For the piezoresistive phenomenon to take place, conductive ma-
terials, such as carbon fibers, carbon black or carbon nanotubes, have to
be provided into the cement matrix. These materials enhance the
electrical conductivity of the cement composites by a magnitude of
order two [71]. Carbon fibers are economic with improved conductivity
and mechanical properties and have proved to be useful for monitoring
of traffic and cracks in asphalt pavement [72]. Carbon nanofibers
(CNTs), which possess good mechanical properties and has good aspect
ratio, can be used as reinforcing the cement sensors [73].
Oven dried CBSS, with 0.5% brass coated steel fibers, was used by
Sun et al. [70] for the study of UHSC columns under cyclic loading of
60 MPa and 120 MPa. The compressive strength obtained for CBSS was
greater than 120 MPa and less noise was recorded by piezoresistivity.
The reason for using the oven – dried CBSS is because of its high sen-
sitivity, making it easier to calibrate than the air-dried CBSS (Table 2).
The test results show a good relationship as a strain gauge (Figs. 4 and
5). It the load case of 120 MPa, a difference between stress and ΔR/ R 0
can be observed at the bottom region (Fig. 5b). This irreversible change
in resistance indicates formation of micro cracks on the specimen.
Therefore proper monitoring of UHSC can be performed using CBSS.
Fig. 3. Piezoresistive behaviour of grapheme based sensor under: (a) successive Azhari and Banthia [69] conducted experiments on the cement
tensile load, (b) under incremental cyclic load [60].
based strain sensors with carbon fibers (CF) and carbon nanofibers as
the conductive material and compared the result with ordinary 10 mm
that during first two cycles of loading and unloading, the resistance foil strain sensors for the cyclic compressive load. The data measured
increased and decreased linearly. But in the third cycle, the resistivity for the sensors were converted to electrical resistivity (Eq. (5):-
did not return to the original value. The irreversible nature indicates
A
damage to the structure, i.e. the graphene epoxy structure has been ρ=R
l (5)
damaged and hence the sensor reading is irreversible in nature.
The sensor used had a gauge factor of 45 with Young’s Modulus and where ρ is the electrical resistivity (Ω cm), A is the area of the cross-
tensile strength of 2.2 ± 0.2 GPa and 80 ± 3 MPa respectively. These section (cm2), R is the electrical resistance (Ω) and l is the length be-
tween inner electrodes (cm).

Table 2
Basic properties of the CBSS [70].
Compressive Strength (MPa) Elastic Modulus (GPa) Poisson’s Ratio Resistivity (Ω m) Gauge Factor

Air-drying 122 ± 4 41.8 ± 2.0 0.17 ± 0.01 (1.41 ± 0.10) * 103 42 ± 10

Oven-drying 134 ± 3 40.7 ± 2.2 0.17 ± 0.01 (3.34 ± 0.32) * 103 202 ± 13

72
S. Das, P. Saha Measurement 129 (2018) 68–90

sensor data obtained from the structure [75]. The data obtained
through BIM can be stored as a digital file and utilized by several ap-
plications [78]. This technology utilizes breakage-triggered (BT) strain
sensors to detect the critical strain [79]. The BT sensors along with
RFID tag can be utilized for wireless communication with BIM-based
SHM software. The BT sensors are connected to the structure along with
a crack-sensitive brittle bar with RFID tag. RFID antenna or reader is
used to scan the RFID tag from a great distance and hence the damage
location and information can be recorded. The RFID tag is activated
only when the pre-set strain is surpassed in the brittle bar [75].
The use of wireless sensing technology along with Micro-Electro-
Mechanical System (MEMS) has proved to be economical and effective
and has been considered as a substitute of the conventional sensors
[80].
Huynh et al. [81] used wireless sensors and implemented Stochastic
Fig. 4. Results from cyclic compression test for stress 60 MPa: (a) Strain vs Subspace Identification (SSI) analysis and Time-frequency Short Time
Time, (b) Stress and ΔR/R0 vs time [70].
Fourier Transform (STFT) method to analyze and update the structural
parameters. The sensors were installed on a bridge and from the ac-
celeration recorded, the modal parameters were found out. It was ob-
served that with an increase in wind speed, the natural frequencies
decreased due to the increment of flexibility and loosening of cables
[82]. The modal parameters obtained were analyzed and updated using
SSI analysis and STFT method [83,84]. It was observed that the STFT
method produced more significant wind inducted variations than SSI.
Therefore, the STFT method can more suitable for computation of dy-
namic response.
The power source of the wireless sensors has to be recharged peri-
odically. To avoid this process, solar panels and rechargeable battery
were attached to each sensor for long-term monitoring process without
any hindrance [81]. The system was a success during the sunny time.
But at night or cloudy day, the power source is cut off thus compro-
mising the system readings. Attaching solar panels to the sensors placed
at the interior points of the structure was costly and complex [85]. Use
of thermal energy was studied for an effective power source [86]. This
Fig. 5. Results from cyclic compression test for stress 120 MPa: (a) Strain vs method uses the temperature difference between the atmospheric air
Time, (b) Stress and ΔR/R0 vs time [70].
and bridge surface to harvest the power for the sensor. However, the
temperature variation may be small which can hinder the power har-
To apply the piezoresistive quality of CF cement based sensors, the vesting process. Another process investigated by the researchers is the
load applied was compared with the fractional change in resistivity radio frequency (RF) harvesting [87]. This method utilizes the radio
(FCR) in Eq. (6):- frequency available in the atmosphere to recharge the sensor power
ρt −ρ0 source. In order to achieve this condition, the RF transmitters have to in
FCR = close vicinity of the RF harvesters, in order to avoid rapid attenuation of
ρ0 (6)
transmission signals [86]. To avoid all these drawbacks, the mechanical
where ρ0 and ρt are the electrical resistivity before the test and at time t or vibrational energy was considered for recharging the sensors. This
during the test respectively. method proved to be effective in case of bridges, in which the power
The samples were tested for load amplitude of 30 kN and load rate was harvested during the high traffic and stored for later use [85]. This
of 0.06 kN/s. The response of the sensor data points was measured for feat was achieved with the help of resonant vibration harvesters, where
ordinary strain sensor and hybrid sensor. It was observed that the error the resonant frequency of the harvester at the sensor point was adjusted
margin improved while using the hybrid sensor. to the modal frequency of the bridge [88–90]. This method cannot be
broadly applied to the bridge as the resonant frequency varies from
3.5. Wireless strain sensor deck to deck [86]. Attachment of automatic tuning devices to the
harvesters may also be used at the cost of a large amount of power or
The sensors normally used for monitoring are connected the central energy [85]. A piezoelectric based self – powered wireless sensor was
server via cable. This process is costly and may provide hindrance in developed to counter the above problems [91]. This sensor uses a series
data acquisition due to the surrounding conditions [74]. In buildings of memory cells to record the load history profile transferred to data
with complex designs, some structural members may remain in- logger system through piezoelectric effect as a voltage signal, which
accessible [75]. Also, the huge amount of data generated has to be charges the sensor [92]. But the major drawback noted was that a part
arranged accurately which is a costly and challenging process [12]. of data is lost in above process which may give insufficient result [93].
With advancements in SHM techniques, building information
modeling (BIM) along with radio frequency identification (RFID) based 3.6. Acoustic emission sensor
wireless structural condition assessment (WSCA) can be employed for
such cases [76]. BIM is a universal virtual user interference and com- With increasing importance of monitoring civil engineering struc-
puting system adopted for SHM system, which can produce the struc- tures, new technology has been developed, acoustic emission (AE),
tural data model of buildings and structures [77]. This system provides which can work under earthquake load of different frequency (20 kHz
good results if it is adopted from the project design and construction to 1 MHz). Acoustic Emission (AE) Method is an NDE method which
phase of a structure. BIM can capture, organize, process and display the uses highly sensitive AE sensors to measure the change in strain, mainly

73
S. Das, P. Saha Measurement 129 (2018) 68–90

useful for detecting crack growth, detection of crack formation, corro-

sion, deformation and void formation [94]. This method works on the
principle of detecting elastic waves due to changes in strain energy
released due to change in structural parameters [95].
Though the method has been successfully installed for different
structure; pipelines, bridges, and others, it faces problem due to at-
mospheric interferences, noise interference [96,97]. Resonant type AE
sensors having a narrow bandwidth can be used for such purposes. This
type of sensors reduces the background noise created due to low fre-
quencies and increases the accuracy of damage localization. The data
acquisition for AE sensor is based on the principle of the threshold; i.e.
if the signal level recorded by the sensor is greater than the pre-set
threshold level, then the AE data is recorded by the sensor [97]. This
data collected is the change in elastic waves occurring due to the newly
formed damage. AE signals have high-frequency range having low
amplitude. The AE signals are in form of the waveform and are recorded Fig. 7. Behaviour of energy b-value with increase in Peak Acceleration [100].
using techniques such as Wavelet Transform (WT) or Continuous Wa-
velet Transform (CWT) [98]. Single AE sensor approach may also detect below 1 indicates severe damage to the structure, which has been cross-
the elastic waves from secondary sources occurring due to friction. In checked by visual observation.
order to reduce this unwanted noise, AE sensors along with load-based Yapar et al. [101] studied the piezoelectric AE sensors for SHM on
or strain based sensors can be used [99]. The use of AE based Micro- bridges and applied mathematical tools like artificial neural networks,
Electro-Mechanical sensors (MEMS) along with strain sensor has been wavelet transformation, and fast Fourier transformation to eliminate
very commonly used in the detection of initiation and growth of cracks problems related to the prediction of the source, noise elimination, and
without the influence of friction [97]. source characterization respectively. The results obtained were com-
Godinez-Azcuaga [99] performed a comparative study between pared with the results of theoretical AE signal measured by Moore et al
wired and wireless AE sensor. The cumulative energy for both the [102] (Eq. (8)).
sensors have been recorded (Fig. 6). It can be clearly seen that the
V = V0 e (−Bt ) sinωt (8)
energy required for the wired sensor was much greater than its wireless
counterpart. The jump in the curves indicates the points where a where V0, V, B, ω, and t are input signal amplitude, output voltage,
sudden growth of damage has occurred. decay constant, angular frequency and time.
AE for damage detection was used to detect the local damage in The experiment was conducted to study steel beam and reinforced
connections to a building under earthquake loads [100]. The data was concrete beam and their comparative study with the predicted signal.
collected using low-frequency sensor of band 25–80 kHz and the digital The experimental results showed good agreement with the predicted
filter was used to reduce the undesirable high and low frequency. For signal having an average error of 1.97% and 1.574% respectively. But
seismic simulation, the noises considered to be present are mechanical for practical use, the number of sensors employed has to be large in
noise from oil flow in the actuator, friction noise from connections order to get accurate damage characteristics.
which were minimized by adjustment of sensors. The damage index Saboonchi et al. [97] made use of capacitive MEMS AE sensors
introduced uses the energy from acoustic signals (energy b-value (b∊)) along with piezoresistive MEMS strain sensor and compared the da-
instead of amplitude. mage detected with respect to piezoelectric AE sensor.
The capacitive MEMS AE sensor consists of two parallel plates
log10 N(AEE) = a−b ∊ log10 (AEE) (7)
acting as electrodes and separated by a dielectric material. The change
where AEE is the true energy of reconstructed signals, a is empirical in capacitance is calculated from the Eq. (9) [103,104]:-
constant and N(AEE) is the total number of signals reconstructed signals δC δε δA δd
with energy higher than AEE. Reconstructed signals are used for cal- = + −
C ε A d (9)
culating AEE; therefore the damage index calculated has little effect on A
secondary source and primary effect of concrete cracking. The sensor where, C = capacitance = ε0 εr d
for interior connection and the exterior connection was tested for ε0 is electric constant ≈ 8.854*10-12 Fm-1, εr is relative static per-
analysis (Fig. 7). It can be observed that the energy b-value reduces mittivity of material between the plates, A is the area of plates facing
with the increase of peak acceleration. The decrease of energy b-value each other, d is the gap between the plates.
Apart from capacitance, the other critical parameter considered for
the design of MEMS sensor is the fundamental frequency of the sensor
(Eq. (10)

Kmech
f = 2π
m (10)
where f is the frequency, m is the mass of the system and Kmech is the
mechanical stiffness.
The piezoresistive MEMS strain sensor works on the principle of
resistance change Eq. (11):-
δR δp δl δA
= + −
R p l A (11)
where ρ is the resistivity coefficient, A is cross-sectional area of the
resistor and L is the length.
The MEMS sensors encounter a problem related to higher level of
Fig. 6. Cumulative energy for Wired and Wireless AE sensor [99]. drift due to the influence of atmospheric thermal conditions [105]. This

74
S. Das, P. Saha Measurement 129 (2018) 68–90

problem can be resolved by the use of the narrow window of sensor Monitoring of bridge with help of accelerometer has proved to be
response [97]. difficult due to low vibrational frequency, lower than 0.2 Hz [116].
Most of the accelerometers are unable to detect this low vibrational
4. Dynamic parameters measurement for SHM frequency [117]. In the recent time, wireless GPS sensors have been
used to in place of cabled sensors and the measurements are taken
The continuous and large structures, bridges and towers, mainly under a clear sky in order to minimize the error due to obstruction.
attract vertical forces, wind force, and have to be analyzed in the fre- Handheld GPS has been used for kinematic point positioning. Also, dual
quency domain for safety against dynamic loading [30]. In order to frequency GPS receivers were used for damage detection. But these
monitor the structures, the acceleration of the structure is monitored sensors faced problem to monitor the damage for various cases; a) sa-
from which the frequency can be obtained. In order to achieve such tellite visibility, b) incorrect tropospheric delay, c) higher sampling rate
condition, accelerometers have been used for diagnostics of bridges and for higher frequencies, 20 Hz [118]. In the recent advancements, the
tall structures. problem with sampling rate was sorted out by using GPS receivers of
sampling rate 100 Hz and improving the signal collection [119,120].
4.1. Accelerometers The monitoring of high rise structures with the help of GPS can be
carried out with the help of RTK technology. This method uses a sta-
Much like the strain sensors, these sensors are the fundamental tionary checkpoint as reference station whose coordinates are pre-de-
building blocks of an SHM system which can detect structural response termined. This station records the difference between its known and its
and use the information for structural diagnostics [106]. A huge calculated position from the data provided by satellites. The difference
amount of data is generated by the acceleration sensors which increases in readings indicates an error which is sent to the rover. Rover is the
the storage cost for the SHM system. In order to overcome this diffi- GPS receiver whose position is tracked and the error information is used
culty, data compression methods have been developed commonly by the rover to improve the accuracy. If the two receivers are close
named as compressive sampling (CS). Bao et al. [107] studied this enough, the error due to clock offset and atmospheric propagation
system and concluded that CS using wavelet basis proved to be more delays can be ignored as errors are strongly correlated. GPS receivers
effective than the Fourier basis system. In order to directly obtain the can be used both for the static and dynamic status of the structure. The
compressed data from the structure, various devices have been devel- static status can be determined using short baseline test (SBL) and zero
oped; Compressive wireless sensing [108], CS camera [109], CS analog- baseline test (ZBL) [121]. SBL is performed with help of two antennas
to-digital converter (ADC) [110]. having receivers on each of them, whereas ZBL is performed with the
help of single antenna with an antenna splitter and two GPS receivers.
4.1.1. Wireless accelerometers For dynamic analysis, the GPS receivers have faced difficulties due to
In order to detect a dynamic response, traditional electronic accel- various effects such as data sampling rate, satellite coverage, multi-path
erometers were used for a long period of time. The accelerometers can effects, receiver noise, etc.
measure three-way acceleration of a structure and are economic and Multipath error occurs when the GPS receiver receives another
can be easily installed [111]. These sensors are known to consume a lot signal from a duplicate satellite or a reflecting surface, thus inducing
of energy and limited resolution capacity [112]. To overcome these error in measurement [122]. A test up was made for GPS based mon-
drawbacks, the wireless monitoring system has been used to which are itoring system and study the multipath effect in a controlled environ-
economic and reliable for remote monitoring within the range of the ment using particle filter algorithm. It was concluded that the method
sensor [111]. was able to reduce the multipath effect significantly for the close re-
Kilic and Unluturk [112] have successfully validated the deflections flector.
occurring in a wind turbine with the help of wireless sensor network Differential GPS (DGPS) is an advanced instrument that has been
(WSN) and conventional accelerometers. The wireless monitoring pro- used for the kinematic point positioning or real-time kinematic (RTK)
cess has proved to be effective for determining the threshold for wind measure. The error due to signal transmission is countered as DGPS uses
turbine strength. Wireless sensors have the capability to detect the total two GPS instruments, one acts as the stable point (base) and the other is
vibrational frequencies generated by the wind turbine as accurately as moved to obtain the different points with respect to the stable point
the ordinary accelerometers [113]. These sensors have to be recharged (rover receivers). The accuracy of DGPS measurement was found to be
periodically and they have to be in the range of computing system. 19.2% for horizontal movement and 28.78% for vertical movement
Wireless sensor technology measurement was validated using [69]. The typical error accuracy for moving points of the structure was
wireless and cable-based accelerometer SHM system on a highway found to be 15 mm in the horizontal direction and 25 mm in vertical
bridge [114]. The wireless and wired accelerometers used were Silicon direction [122–124].
Designs 2012 accelerometer and Honeywell QA750 accelerometer re- For monitoring of dynamic loads such as an earthquake, high-fre-
spectively with sampling rate 100 Hz and 256 Hz respectively. The quency sampling needs to be used. DGPS needs the base station to be
sensors were placed in 19 locations at the same level along the bridge stable for good data analysis. Therefore in case of an earthquake, this
with a maximum distance of 40 m. The 5 s acceleration data obtained sensor is not suitable for measurement as the movement of the base
for location 6 and 10 has been shown in Fig. 8. It can be seen that the receiver may include the noisy data [125]. In light of this event, Precise
readings of both the sensor system totally match, indicating that Point Positioning (PPP) is a new technique that has been developed.
wireless sensor system can be used for short-term SHM. The wireless Precise Point Positioning (PPP) is a new technique for accurate posi-
sensor was packed in a waterproof box along with battery pack. tioning for dynamic applications [126]. The accuracy of PPP is around
But these sensors come across interference from electromagnetic 2 cm in the horizontal direction [125]. This technique makes use of a
radiations, thus reducing the accuracy. To reduce this phenomenon, moving receiver and the corrections are made using the data collected
further research in this field has been conducted [112]. from a relatively large number of remote receivers operating perma-
nently [127]. The limitation of GPS sensors is that a bulk number of
4.1.2. GPS based accelerometer observations have to be collected for high-frequency sampling. The use
The geodetic monitoring system developed for monitoring of of GNSS-PPP using high rate GPS can be used to overcome this lim-
structure uses mainly global navigation satellite systems (GNSS) or itation [128].
global positioning system (GPS) [115]. This system measures the po- In order to avoid the other inconsistent problems and to increase
sitions of monitored points in three directions for both dynamic and reliability and accuracy, an integrated sensor system was introduced.
static cases and monitoring is done in frequency and time domain. This system may contain GPS receiver, strain gauge, displacement

75
S. Das, P. Saha Measurement 129 (2018) 68–90

Fig. 8. 5s acceleration data obtained from cabled and wireless accelerometer for: (a) location 6 and (b) location 10 [114].

transducers or accelerometers and has proved to increase the pro- analyse the frequency values of accelerometer and GPS data to obtain
ductivity of the SHM system [129]. For the integrated type sensors, the original vibration frequency of the structure. 7% error between the
extraction of structural dynamic properties has to be performed using a two sensor readings was obtained which was good for damage detec-
proper digital filter or bandpass filter, Chebyshev type digital filter tion.
[130]. For detection of dynamic deformation, sensors such as, high fre-
Meng et al. [118] experimented the integrated sensor consisting of quency GPS/GLONASS receivers, tachymeters, inductive displacement
GPS and triaxial accelerometers on a suspension footbridge. The GPS sensors, code levellers and meteorological station have been set up. The
receivers, sampled at 10 Hz, and triaxial accelerators, sampled at 80 Hz, accuracy of GPS/GLONASS receiver was noted as accurate within 2 cm
were installed along the span of the bridge. The vertical acceleration and the tachymeter and displacement sensor accuracy was recorded to
obtained for damage event for re-sampled acceleration (10 Hz) and GPS be within ± 1 mm. Therefore this integrated GPS sensor system was
are shown in Fig. 9(a) and (b) respectively. It can be observed from effective for SHM [133].
Fig. 9(a) that on the removal of forced excitation, the amplitude of
acceleration reduced to 0 gradually, but the random noise of less than 4.1.3. Micro-electro-mechanical system (MEMS) accelerometer
1 cm amplitude was recorded in GPS measurement (Fig. 9(b)). There- With the improvement in technology, the researchers are focussed
fore GPS sensors proved to be more sensitive in detecting the low fre- on developing Micro-electro-mechanical systems (MEMS) accel-
quencies and hence more effective than the normal accelerometer. erometers based on the working principles of capacitive, piezoelectric
Therefore integration of both RTK GPS sensor and triaxial accel- and piezoresistive accelerometers [106,134]. But only capacitive and
erometers can be used for effective damage monitoring, as it can gather piezoresistive accelerometers are used in the modern world. Though
data for low sample rate (10 Hz) and high sample rate (100 Hz) re- piezoresistive accelerometers are more sensitive in high-frequency
spectively[131]. However for the system to work, absolute positioning bands, the capacitive based MEMS accelerometer is popular due to the
of sensors has to be pin pointed in order to use an algorithm to integrate temperature coefficient problem with the former [80,135]. The key
the sensors. Keeping the problem of sensor data integration, sensors parameters for the effective design of an accelerometers sensor are
with high data sampling rates were studied. JNS100 and Leica receivers deflection, stress sensitivity and resonant frequencies [136]. Accel-
showed a good sampling frequency of 50 Hz which is close to the erometers have sensitivity inversely proportional to the square of first
normal triaxial accelerometers (80 Hz). resonant frequency [136,137]. Accelerometers are mainly designed for
The other obstacle faced by the GPS sensor integration system is the sensing high frequency, hence having low sensitivity. But the natural
accelerometer and GPS time series synchronization [132]. To avoid this frequency of a structure is less than 100 Hz which cannot be detected by
mismatch of time series, a frequency based algorithm was devised to the traditional MEMS accelerometers [106,135]. In most of the cases,
the ambient vibrations occurring in the structures have a small value of
g, low amplitude, except for earthquakes having 1–2 g [138]. The
readings of these accelerometers are mainly affected by thermal noise
caused due to piezoresistors and due to thermal agitation in sensing
structure [139]. Therefore MEMS accelerometer with small bandwidth
and simple signal conditioning electronics has been designed to im-
prove the sensitivity as well as the power consumption and has ultra
low floor noise [137]. This technology can be efficiently used for
wireless sensor networks [140].
A study has been performed by Kavitha et al. [136] to compare the
performance of piezoresistive MEMS accelerometers and capacitive
MEMS accelerometers (Fig. 10) with respect to frequency and deflec-
tion sensitivity. It was observed that the capacitive sensor has a re-
sponse range of 0–20 Hz and very low floor noise, hence is suitable for
low-frequency acceleration in SHM as well as seismic applications. A
brief comparison has been made with that of the available Commercial
Off-the-shelf technologies (COTS) accelerometers with the piezoelectric
and capacitive accelerometers [141,142] as shown in Table 3. The ca-
pacitive and piezoresistive MEMS sensors proved to be superior to the
other sensors used.
Fig. 9. Vertical Acceleration data in Event 1 for : (a) Re- sampled acceleration Kavitha et al. [137] designed two ADXL series MEMS accel-
(10 Hz), and (b) GPS accelerometer [118]. erometers with comb finger type differential capacitive MEMS structure

76
S. Das, P. Saha Measurement 129 (2018) 68–90

macrocell as cathode and rebar or macrocell as the anode, and the

macrocell current measurement can be used for evaluation of corrosion
rate [157]. The corrosion rate using macrocell measurements was
successful in proving information about passive and active corrosion.
However in practical implementation, this sensor faced problem due to
interference of environmental wet and dry cycle and electric resistance
to concrete [158]. In order to improve the measurements, steel/copper
galvanic sensor has been used which provides good corrosion rate of
steel bar in concrete [159]. Another sensor configuration is the use of
capacitive transducers (CT) which works on the principle of non contact
based corrosion detection. This sensor uses two capacitance electrodes
on both sides of rebar to detect corrosion as well as the thickness of rust
on rebar due to corrosion [148]. The sensor was effective for thick rebar
and was also able to detect clean steel from corroded steel.
Printed Circuit Board (PCB) based corrosion sensor was designed
Fig. 10. Performance of: (a) Piezoresistive MEMS accelerometer and (b) ca- which can resonant frequency variation of couple coil resonator to
pacitive MEMS accelerometer.
detect corrosion [160]. The difference in resonant frequency can also be
detected using wireless sensors at the early stage of corrosion detection
[143], Device A for low-frequency ambient vibrations and Device B for [161]. Based on the principle of wireless sensing technology, RFID
earthquake applications. The capacitive accelerometers are simple in based corrosion sensors embedded in the concrete have been able to
design and have low drift, good sensitivity and noise performance detect the chloride concentration change in concrete [162]. The sensor
[144]. The performance of both the devices was compared with other can be used for long term using small amount of power from Radio
MEMS accelerometers widely used (Table 4) [145,146]. It was noted Frequency (RF) signals [163].
that Device A provided good noise performance. Device A and Device B
proved to have good voltage sensitivity and excellent noise immunity 6. Multiple damage detection
which makes them suitable for SHM and seismic applications respec-
tively. The techniques discussed above all deal with detection of a parti-
cular type of damage i.e., strain, temperature, acceleration or dis-
5. Corrosion sensors placement, etc. The application of all the sensors at a particular struc-
ture is not feasible. Therefore the need of sensor that can detect
Another type of damage analysis in case of damage failure is due to multiple damages was strongly needed. In light of this event sensors
corrosion in structural rebar steel [147]. In the case of corrosion, the based on fiber optic principle was developed which can be utilized for
first time corrosion detection is done directly by determining the multiple damage detection [33].
thickness of rust layer [148]. The main source of corrosion is due to
chloride contamination [149] and change in pH value [150] inside 6.1. Fiber optic-based sensor
concrete. A number of corrosion detection techniques have been ex-
perimentally invested such as; polarization resistance technique [151], Most of the traditional sensors are attached with the help of cables.
Impedance Spectroscopy [152], galvano-static pulse transient tech- These cables over a long distance may suffer from electromagnetic in-
nique [153]. However these techniques, in the laboratory have been terference (EMI). To overcome this phenomenon, Fiber Optic sensors
tested for gradual corrosion and the area of steel was fixed. In the field have been introduced for SHM. The advantages of Fiber Optic based
these techniques posed a problem as corrosion in real life is not gradual. sensor over traditional sensors are that they are lightweight, corrosion
Embedded sensor such as piezoceramic lead zirconate (PZT) trans- resistant and durable. These sensors are immune to electromagnetic and
ducer, which uses acoustic emission (AE) technique, has been used for radio interference, due to dielectric materials, thus making them sui-
corrosion monitoring of reinforced concrete structure [31,154]. These table for monitoring metal structures even for small length, 2 µm, with
sensors use ultrasonic waves to indicate different stages of rebar cor- high precision [33,164]. The fiber optics sensors are also versatile and
rosion [155]. The embedded sensor can be installed at the time of the same optical fiber sensor can be used for measurement of multiple
construction, therefore corrosion monitoring of existing structures is parameters; humidity, temperature, pressure, displacement, accelera-
not feasible. The surface bonded embedded PZT sensors were compared tion, magnetic flux and pH value. Therefore the need for multiple tra-
with embedded PZT sensors for corrosion detection. The surface bonded ditional sensors is reduced [34]. These sensors are small in size, low
sensor was able to detect progression of sensors and the embedded cost and low power consumption [165]. Optical fiber consists of a) core,
sensor was able to detect early stage of rebar corrosion, mainly the a thin glass fiber, b) cladding, that confines the propagation of light
percentage of stiffness loss [156]. within the fiber core and c) protective coating, which provides me-
Likewise various sensor configurations were used for corrosion de- chanical strength and absorbs moisture (Fig. 11) [166]. The basic ma-
tection. One among them is the galvanic couple, which makes use of terial of this sensor is silica which can withstand high temperature up to

Table 3
Comparison of Capacitive and Piezoresistive accelerometers with COTS accelerometers [141,142].
Device/Manufacturer Range Measured sensitivity Flat frequency response range (Hz) DC – 100 Hz, 0–2 g resolution (μg) Amplifying electronics

Silicon Design (SD2012-10) 0–10 399.3 mV/g 1–200 124 Yes

M1220D (Motorola) 0–8 241.4 mV/g 1–90 1237 Yes
Jerome [97] 0–10 1.250 mV/g 0–680 20 No
ADXL210A (Analog Device) 0–10 108.0 mV/g 1–16 258 Yes
7290A-10 (Endevco) 0–2 199.9 mV/g 1–200 84 Yes
Capacitive accelerometer 0–2 301 mV/g/V 0–30 9.6 No
Piezoresistive accelerometer 0–2 4.0 mV/g/V 0–45 12.72 No

77
S. Das, P. Saha Measurement 129 (2018) 68–90

Table 4
Comparison of two ADXL series accelerometers with other MEMS accelerometers [137].
Device/Manufacturer Range (g) Specified sensitivity Resonance frequency (Hz) Noise floor (μg/ Hz ) Micro machining process MEMS type

Colibrys (MS9002) 0–2 1000 (mV/g) 100 18 BM Silicon

Endevco (7290A) 1000 (mV/g) 1300 BM Silicon
Silicon Designs, INC. (1221) 0–2 2 (V/g) 400 5 BM Non-Silicon Nickel
Kavitha et al. [138] 0–2 4 (mV/g/V) 100 4.53 BM Silicon
Li and Tseng [139] 0–2 0.40 (V/g) 500 1.6 BM and SM Silicon
Device A 0–0.1 1915 (mV/g/V) 100 1.3 SM Polysilicon
Device B 0–2 106.8 (mV/g/V) 500 4.64 SM Polysilicon

mechanically attached and welding [173]. The process of welding of

sensors to the structure can alter the chemical properties. Therefore this
process is not favoured in the practical field [174].
The major disadvantage of these sensors is that they are fragile
[175]. Being fragile, the FBG sensor has to be encapsulated properly
before installing it on structural element [176]. The protective system
provided for the FBG should be such that it can: a) withstand harsh
environment during the construction stage, b) isolation of the sensor so
that it might withstand physical shock or pressure and keep the sensor
intact, c) in-penetrable to moisture and d) resistant to chemical attacks.
Ngoi et al. [177] used cylindrical silicone rubber tube to embed the
Fig. 11. Single mode optical fiber. FBG, which proved to be effective in detecting lateral loads. The sili-
cone rubber tube packed FBG sensor has high Poisson’s ratio, low
elastic modulus and is thermally stable in the range −100 to 320 °C
1000 °C and low temperature of less than 4 K, suitable for vacuum
[171]. Dawood et al. [178] used vacuum infusion technique to embed
condition with good fiber coating and resistant against radiation as the
the FBG sensor between cross-ply laminate of GFRP sandwich material
core of pure silica can stabilize the change of refractive index due to
and foam core. This arrangement is useful for microscopic sensing of
radiation [167]. Three technologies of fiber optics present are; point
localized defects at an early stage under dynamic and static loading
sensor (Fiber Bragg Grating (FBG) sensor) [168], long gauge sensor or
with great accuracy. Chung and Kang [179] used a technique to embed
quasi-distributed sensor (multiplexed Fiber Bragg Grating sensors)
FBG in steel rebar and bond it with adhesive to the rebar. This is an easy
[169] and fully-distributed sensor (Brillouin Optical Time Domain Re-
technique than vacuum infusion but the accuracy of strain measure-
flectometry (BOTDR) and Brillouin Optical Time Domain Analysis
ment has to be verified.
(BOTDA)) [170]. Out of these sensors, point sensor; Fiber Bragg Grating
The FBG sensors have to be attached to the structures with the help
(FBG) sensor, is mostly used for practical monitoring of structures
of adequate adhesives, failing to which errors due to temperature, in-
[171].
sensitivity and loosening might occur [180]. For long-term monitoring,
adhesives with higher service life such as epoxy and acrylic type ad-
6.1.1. FBG sensor
hesives are preferred. For short-term and dynamic monitoring, fast cure
The Bragg gratings operate by acting as wavelength selective filter
adhesives like instant cyanoacrylates adhesives are used [172]. The
that can reflect single particular wavelength, Bragg wavelength, and
adhesives commonly used to attach the sensor to the structure are cy-
transmit the other wavelengths (Fig. 12) [166]. For the FBG to function,
anoacrylate C2 and epoxy NM BPE Lim 465 [181]. It has been experi-
it is illuminated by a bright light source and the measurements are
mentally observed with the help of modified fibre pull-out test that
taken due to the change in wavelength of reflected light [172].
cyanoacrylate can withstand higher failure load than epoxy glue [181].
The Bragg wavelength is given as:
It was observed by Torres et al. [166] that the presence of packaging
λB = 2Λneff (12) material below the sensor can cause errors in reading. For new struc-
tures, the FBG can be embedded into the structures during construction
where neff is the mean refractive index of the core and Λ is the grating
phase without any alteration in structural properties [171].
pitch.
Moyo et al. [175] compared the tensile test of steel rebar and static
The installation of FBG to a structure is a major hindrance. They can
test results of concrete beams for strain measurement between FBG
be attached to a structure by the process of adhesive bonding,
sensors and electrical resistance strain gauges (ERS). The FBG sensor
was packed between two layers of carbon composite material. The
tensile test result shows good correlation between the strain measure-
ments of two sensors. The static tests of the concrete beam were con-
ducted on top and bottom parts at mid-point and the results show good
linearity between the two sensors. The system had a sensitivity of 1 με
and accuracy of 5 με. Thus it was noted that FBG can be used for ef-
fective strain measurement. For dynamic loading, the ERS and FBG
showed a maximum strain of 58 με and 55 με respectively.
For effective diagnostic of structure, the sensor used should be
sensitive to only one parameter. But a bare FBG sensor is sensitive to
pressure, strain and temperature. Therefore while measuring strain; it
may also detect strain due to thermal variations [175]. In order to
minimize this cross-sensitivity, measurement at two different optical
modes for which temperature and strain responses are different [182].
Fig. 12. Transmission and Reflection Spectra from a FBG sensor [166]. A method was developed to measure the strain and temperature with

78
S. Das, P. Saha Measurement 129 (2018) 68–90

the help of thermochromic material and FBG sensor. The variation of multi-purpose sensor such as FBG sensor can be used for analysis owing
strain was measured with the help of FBG sensor, and the temperature to its good strain sensing and easy surface mountable property. The
was measured by the change in optical power reflected by the ther- measurement result of the FBG system showed good agreement with the
mochromic material. This complex technique showed good results but dynamic strain recorded on the traditional resistive strain gauge over
limited measurement range [183]. variable temperature changes [191]. Similar study was conducted on
The change in strain and temperature can be identified with the Lezíria Bridge in order to validate the FBG sensor response with con-
change in wavelength of the FGB [184]: ventional strain sensors and FBG proved to be a more robust for strain
detection [192].
ΔλB
= Pe εzz + [Pe (αs−αf ) + ζ ]ΔT Newly designed SHM module consisting of wireless sensor network
λB (13)
was used on the Jiubao Bridge [193]. The sensor module included
where αs and αf are coefficients of thermal expansion of structural several self – devised FBG sensors which were used for strain, tension,
material and optical fibre respectively, ζ is the thermo-optic coefficient, and temperature measurement. Other sensors such as an accelerometer,
εzz is the axial strain and Pe is strain-optic coefficient. displacement transducers were also a part of the sensor system. The
The effects of temperature, strain and pressure can be modelled by data was transmitted to the workstation via the internet. The strain
the following equation [175]: sensors are embedded in the deck and placed on the surface for mea-
surement. Temperature sensors were placed steel girder, concrete slabs,
ΔλB = K ε ε + KT ΔT + KP P (14)
and arches. The temperature at the middle span of bridge deck was
where Kε, KP and KT are the coefficients of wavelength sensitivity to monitored. It was seen that a distinct difference in temperature was
strain, pressure and temperature respectively [171]. observed for top and bottom deck, the top deck being at a lower tem-
perature than the bottom, thus measuring the temperature change in
K ε = [1−0.5neff (ρ12−ν (ρ11−ρ12))] λB
the bridge. The strain data was collected from the FBG strain gauge and
the data were combined into average power spectral density. The
KT = [α + ξ ] λB
identified frequency was similar to the results obtained from the 3D FE
1−2ν n2 model. This result verifies the reliability of the monitoring system.
KP = ⎡− + (1−2ν )(2ρ12 + ρ11) ⎤ λB The monitoring of bridge was performed using distributed fiber
⎢
⎣ E 2E ⎥
⎦
optical sensors. These sensors are good for measurement of strain in the
where ρ11 and ρ12 are components of fiber optic strain tensor, n is the bridge and measure the change in stiffness for damage identification,
refractive index of fiber, ξ is fiber thermo-optic coefficient, α is the locating the damage and measure the extent of damage [194]. Strain
thermal expansion, ν is Poisson’s ratio of fiber and E is Young’s Mod- influence line theory was the primary proposed method for vehicle load
ulus. For wavelength of 1.55 μm, parameters Ke = 1.15 pm/με, identification, however, the traditional strain gauge was able to obtain
KP = −3.0 pm/MPa and KT = 11 pm/0C remain constant for bare FBG. readings for local damage [195]. The method was tested on a model
But the sensitivity coefficients may vary for different OFBGS. Therefore bridge having a biaxial model vehicle load and long gauge FBG sensors
calibration has been to do for individual sensors before installing [185]. for measuring strain. The long gauge FBG sensor is made to prevent
It has been observed that for 1 °C temperature shift, change in Bragg from the harsh environment with the help of basalt-fiber-reinforced
wavelength is approximately 10 με. polymer. The model was tested for damage identification both for the
For the measurement of both strain and temperature, a new tech- analytical model and experimental model. The location and extent of
nique was introduced by the use of optical Fiber Bragg Grating Sensors the damaged element were identified with the help of long gauge FBG
(OFBGS). In order to protect the bare OFBGS against breaking while sensors and the result was not affected due to the velocity of the ve-
installation, they are encapsulated in glass fiber reinforced polymer hicle. As the initial condition of the bridge is not known, equivalent
(GFRP) [186]. With the integration of GFRP bar with OFBGS, the sen- damage extent is implemented to assess the condition of the bridge.
sing properties, as well as the mechanical properties of the sensors, This value is based on the average value of long gauge stiffness coef-
were found to increase. ficient of intact elements. The other two damage extent defined are
The use of distributed FBG sensor in case of bridge monitoring has designed damage extent, based on the variation of the moment of in-
been very common. In case of bridge monitoring, a number of algo- ertia and ideal identified damage extent, which is based on the stiffness
rithms have been developed to check the status of the bridge under coefficient of intact state. For double damage detection, equivalent
heavy vehicular loads. These methods make use of strain sensors for damage extent is considered as the reference to evaluate the condition
measuring the change in strain and then processing the measured data of the bridge. The experimental and analytical result for the damage
for damage analysis [187]. However, the analysis of data for sensors far coincides to prove that the proposed long gauge FBG sensors are ef-
away from the damaged area gave false results. The use temperature fective for damage detection.
induced strain from the measured strain data to detect an abnormality
in the bridge was used [188]. A Euclidean distance-based damage index 6.1.2. Fiber optics sensor
method is defined which uses temperature variation and temperature Fiber Optics Systems (FOSs) use fibers as the signal transmission
induced strains to detect damage in the structure. This method can be medium and sensing element over long distances [171]. The FOSs in-
used for long-term monitoring of bridges as stiffness reduction for cludes Fabry-Perot (FP), Surveillance d’Ouvrange par Fibre Optics
multiple damage locations can be measured. However, the limitation of (SOFO), and Brillouin sensors [174]. These FOSs are flexible, multiplex,
this method was that the visual inspections need to be performed for versatile and cost-effective [196].
damage of other types as a reduction in stiffness does not occur due to The FP sensor is an interferometer which uses multi-mode fiber, a
damages such as corrosion. simple and economic technology, and can be made temperature self-
In order to develop a method for damage location in multiple lo- compensated by taking into account the thermal expansion of the host
cations of a structure, the Wind And Structural Health Monitoring material. Therefore the strain readings recorded by the sensor is only
System (WASHMS) was used for study in Tsing Ma Bridge [189]. This due to stress-induced [197]. This sensor uses Michelson fiber-optic in-
system uses eight types of sensors such as temperature sensor, dynamic terferometer arrangement in which one fiber is prestrained (0.5%) and
strain gauge, GPS, servo-type accelerometer, anemometer, level sensing attached to the host structure to act as a sensing arm. The other fiber,
stations, displacement transducers and dynamic weigh-in-motion sta- reference fiber, is laid close to the previous fiber parallelly so that the
tions for continuous monitoring of loads, structural responses and en- temperature variation between the two is negligible [196]. Both the
vironmental factors [190]. In order to reduce the number the sensors, fibers are installed in a tube. The difference in length of the two fibers

79
S. Das, P. Saha Measurement 129 (2018) 68–90

gives the measurement of strain induced due to compression or tension

[198]. SOFO system makes use of an additional broad light source, LED,
for damage identification [174]. This configuration is known as low-
coherence double Michelson interferometer [199]. The first inter-
ferometer is the pair of single mode fibers as the previous sensor. The
second interferometer consists of a scanning mirror which with the help
of light source (1.3 µm of LED) can detect the difference in length be-
tween the two fibers. The stability and precision obtained by this sensor
is 2 µm [199]. The data acquisition is performed at an interval of 7 s
and the dynamic range of measurement is from −0.5% to 1% [164].
Brillouin sensors, BOTDR AND BOTDA, uses Brillouin scattering which
is produced due to non-linear interaction of acoustic waves and light.
Brillouin Scattering effect may also be defined as the phenomenon
which is stimulated when there is a frequency difference between pump
pulse light and wave light propagating in the FOS matches with the Fig. 13. Graphical interpretation of the theoretical and experimental results
local Brillouin frequency of fiber core [200]. The relation between light [164].
frequency and strain is given by Eq. (15) [201]:
ν (ε ) = ν (0)(1 + Cε ∗ε ) (15) Antunes et al. [30] used this technique for SHM of a tower for bi-
axial sensing. The discrepancies or the difference in frequencies ob-
where ν(ε) is Brillouin scattering frequency due to change in strain, tained are due to the overestimation of stiffness or mass distribution in
Cε is a constant coefficient due to change in strain, ν(0) is the reference numerical process. Therefore, the FBG biaxial accelerometer can be
frequency. used effectively for SHM process.
This phenomenon induces small dynamic change, due to change in
strain and temperature, in the fiber core refractive index [202].
6.1.4. FBG as corrosion sensor
Therefore, the Brillouin frequency can measure both strain and tem-
Most of the corrosion sensing techniques require direct contact with
perature along the fiber. The only drawback of using FOSs is that they
the steel rebar for detection which is always not possible. Fiber Optics
are fragile and has to be installed with extra care [174].
Sensors have been advantageous for corrosion detection due to its low
The most economic arrangement of FOS for a bridge is tree pattern
cost, high resistance to fracture and easy termination [204]. These
with one main cable along with different branches of the optic fiber.
sensors can be embedded easily with good protective cover and the
This arrangement also helps in multiplexing as the system can in-
sensor changes its fluorescence intensity with the change in pH [205].
corporate a large number of sensors using only one demodulation
These sensors, being immune to electromagnetic interferences, have got
system [174].
Uva et al. [164] made use of SOFO sensors to evaluate the health an upper hand to the normal embedded sensors which are susceptible to
external electric field and temperature fluctuations [206].
condition of the beams of a bridge. Each beam was fitted with 6 SOFO
The FBG sensor for corrosion detection has been tested for without
sensors coupled with 6 thermocouples. The thermocouples were in-
any coating and with Polydimethyl Siloxane (PDMS) coating [207]. The
stalled to evaluate the change in strain due to temperature change in
change in strain due to corrosion is sensed by the FBG sensor. The
the beams. The experimental value of strain variation was calculated
sensors were embedded in concrete for testing. The FBG sensor with
for time t1 and t2 from Eqs. (16) and (17) respectively, taking the length
PDMS coating was observed to detect corrosion at a more early stage
of measurement fiber LS as 300 mm.
than that of the uncoated FBG. The PDMS coating being soft and thin,
ms (1)−ms (0) acts as a media for transfer of corrosion strain to FBG sensor for de-
ΔεEXP (t1) =
LS (16) tection [208]. Whereas, the uncoated FBG sensor needs the rust to be in
close contact with the fiber so that wavelength shift is detected due to
ms (2)−ms (0)
ΔεEXP (t2) = early corrosion. However the PDMS layer has a low Young’s Modulus
LS (17) which may affect detection in long term [209]. Also the PDMS layer
where ms(0) is the strain measured at the launching of the beam in time may absorb chemicals in concrete, which may affect the damage de-
t0, ms(1) is the strain measured at time t1 after completion of slab and tection [207]. To avoid this reaction of the FBG coating layer, a pH-
ms(2) is strain measured at time t2 after completion of the bridge. The sensitive hydrogel was mixed with the PDMS coating [210]. This
graphical interpretation of the theoretical and experimental results has coating is used mainly in humid region where the hydrogel layer is pH
been shown in Fig. 13. The structural integrity is obtained from the sensitive and PDMS is strain sensitive [208,211]. The mixed type coated
comparison between the two curves. The red zone shows the data above FBG sensor showed monotonic relation between wavelet shift and
the predicted strain which indicates an unaccounted phenomenon, i.e. corrosion process which makes it a robust corrosion detecting sensor.
the strain level is above the predicted safety zone. The comparative study of different types of strain, acceleration,
Hong et al. [201] used the BOTDA based FOS to study the strain corrosion and FBG sensors depending on their type, material used,
inducted in H-piles during pile driving process. The experimentally damage detection type, working principle, advantages and limitations
measured strain data was verified with the theoretical model developed have been tabulated in Table 5.
by Misra and Chen [203]. It was observed that most of the data col-
lected from the FOS match with the predicted data. Therefore the 7. A Comparative case study among different sensors
BOTDA based FOS can be adopted for practical use.
7.1. Strain sensors
6.1.3. FBG based accelerometer
In order to overcome the drawbacks of normal or wireless accel- It has been observed that normal metallic thin foil strain gauges are
erometers, advanced Fiber Bragg Grating (FBG) based accelerometers unable to measure dynamic strain more than 2000με for 106 alternating
have been introduced. These sensors provide immunity against elec- load cycles [56]. Introduction of optical strain gauges has shown great
tromagnetic interference, can be monitored from a remote location and potential in the measurement of large dynamic strains. They can resist
is suitable for embedding into the structure [34,171]. large strains and alternating loads [212]. Experimentally it was seen

80
 S. Das, P. Saha

Table 5
Comparative study for the different sensors used for health monitoring of structure.
Sensors Type Material Damage Type Working Principle Advantages Limitations
Detected

Machine Based Vision Technology Digital and Video High Resolution Digital Crack Patterns, Crack Superimposition of damage Non Contact technology, No complex Capturing image at low light, Angle of
Photography Camera Growth templates calculation required for damage detection elevation for the camera, Impaired vision
due to climate
Metallic thin film strain gauge Electrical, Wired Thin Metallic Film Strain Change in Strain Economic, Stable and Compact Dynamic Loads cannot be measured
Thick film ceramic strain sensors Electrical, Wired Thick Piezoresistive Strain Piezoresistive Effect-Voltage Durable, Corrosion Resistant, Elastic Calibration for atmospheric effects has to
material-Ceramic Change be performed
Carbon based graphene sensor Electrical, Wired Carbon Nanomaterials- Strain Change in Electronic Energy Lightweight, Highly Sensitive Can be used for particularly one type of
Graphene Band damage parameter
Cement based strain sensor Electrical, Wired Cement Strain Change in Volume Electric Suitable for Embedding in Concrete Can be used for particularly one type of
Resistivity damage parameter
Wireless strain sensor Electrical, Wireless Electronic Equipment Strain Radio frequency identification Automated System, Can Monitor Remote Costly, Constant Power Source Required
Areas
Acoustic Emission Sensor Electrical, Wired, Piezoelectric material, Strain Change in Elastic Waves Low amplitude, Detect damage at lower Influence of atmospheric temperature
MEMS threshold

81
Wireless Accelerometers Electrical, Wireless Electronic Equipment Acceleration Measure three way acceleration Economic, Remote monitoring. Electromagnetic interference
of a structure
GPS Based Accelerometer Electrical, Wireless GPS Acceleration Change in vibrational frequency Sensitive a low frequency, Remote sensing Power consumption
using a proper digital filter
Micro Electro-Mechanical System Electrical, Wireless Piezoelectric material Strain Acceleration Piezoresistivity and capacitive Low bandwidth, Noise reduction, Low High level of drift due to the influence of
(MEMS) Technology Sensors power consumption, High sensitivity atmospheric thermal conditions
Corrosion Sensors Electrical, Wired, Piezoceramic material, Corrosion (Rusting AE waves, Cathode Anode Can be used as surface embedded (non- The alternate wet and dry environment
Wireless galvanic couple and pH change) arrangement, RF signals contact) sensor for rebar corrosion condition may hamper the sensitivity
detection
FBG Sensor Optical Optic Fiber Multiple Light refractivity Effective for local damage detection at Fragile, hence proper care to be taken while
points installed, Free from installation
Electromagnetic Influence
Fiber Optics Sensor Optical Optic Fiber Multiple Light Refractivity Effective for damage detection for entire Fragile, hence proper care to be taken while
length of fiber, Free from Electromagnetic installation
Influence
FBG based Accelerometer Optical Optic Fiber Acceleration Light Refractivity Immune against electromagnetic influence Fragile, hence proper care to be taken while
installation
FBG based Corrosion Sensor Optical Optic Fiber Corrosion Light Refractivity Immune to external electric field and Fragile and need to be embedded in
temperature fluctuations concrete with protective coating
Measurement 129 (2018) 68–90
S. Das, P. Saha Measurement 129 (2018) 68–90

that optical strain gauges were able to resist alternate loading strain of
magnitude ± 1000 με for 107 loading cycles [213]. But in the practical
field, these sensors were not able to provide a significant advantage
over the traditional sensors for dynamic loading. Also, the sensors are
costly and the power consumption may pose a problem for wireless
monitoring [214].
Neild et al. [67] conducted practical and theoretical experiments to
conclude that for sensor embedded in massive concrete structure and
sensor on the surface, the readings may vary. Sreeshyiarn et al. [215]
introduced temperature calibration factor for the embedded sensor as
2.19 με/0C and for gauge mounted on the surface as 4.32 με/0C.
However, they also expressed the need for individual calibration test for
each VWSG due to the varied thermal expansion of concrete. A com-
parative experimental study was conducted by Ge et al. [53] on an RCC
beam to investigate the effect of temperature change in different strain
sensors; conventional foil electrical resistance strain gauge (ERS), Vi-
brating Wire Strain Gauge (VWSG), FBG sensor and distributed Bril-
louin backscattering optic fibre (FO) sensor. The theoretical strain for
the beam was calculated considering the thermal expansion coefficient
for concrete as 11 με/°C [216]. The strain results obtained from the
different sensors are being compared with the theoretical strain calcu-
lation. Two conditions were maintained during the experiment; (1) the
beam is kept in an insulated chamber with temperature change from
room temperature to 40 °C by increments of 1 °C (Fig. 14); and (2) the
temperature was applied only from above, keeping the bottom part
open to ordinary temperature, thus creating temperature gradient
(Fig. 15(a) – at top and Fig. 15(b) – at bottom). It can be clearly ob-
served that the VWSG has got an upper hand against all the other
sensors for measurement of average strain due to dynamic change in
temperature. ERS proved to be unreliable for this case study. FBG
showed 4% difference the actual strain. This difference can be solved by
calibrating each FBG sensor instead of considering an average tem-
Fig. 15. Strain measurements for temperature gradient case [53].
perature coefficient. As for FO sensors, it was observed that the mea-
surements were inconsistent, having a difference of 5 °C with the the-
oretical value. However, it has been observed that for long-term theoretical strain was calculated using Eq. (18), taking Δp and Δε as 0
monitoring, where the temperature change is not sudden, FO sensors and αM as 10-5 0C-1 [175].
have been effective. ΔλB / λB = Pε . Δε + [Pε . (αM −αF ) + ζ ]. ΔT + GP . Δp (18)
Costa and Figueiras [174] experimented the static behaviour of a
simply supported beam with the help of electric strain gauge (ESG) and where Pε and ζ are the coefficients of strain and thermo-optic of the FO,
fiber optics of length 4 cm and 10 cm. A maximum load of 60kN has GP is the gain factor for pressure and αM and αF are the thermal ex-
applied in a pair 50 cm apart. The FO was protected by polyester sheet pansion coefficients for host material and fibre respectively. The test
and attached with the help of epoxy resin adhesive. In the two cases, the was conducted with the help FOS and resistive thermometer for one
performance of ESG was compared with; (a) FOS of 4 cm length and (b) temperature cycle ranging between 30 °C and 60 °C. It was observed
FOS of 10 cm length (long gauge). Differences in strain observed be- that the theoretical and experimental strains were exactly the same for
tween the two sensors for the first and second case were 20% and 1.5% the steel specimen (Fig. 17). Therefore the FOS along with the protec-
respectively (Fig. 16(a) and (b)). Therefore it was concluded that FOS tive cover provides no hindrance to the measurement of strain in dif-
with 10 cm length can be used for practical use. The researchers also ferent cases.
carried out experiments on a steel specimen for determining the effi- The efficiency of strain measurement for sensors, carbon nano tube
ciency of FOSs in determining strain due to change in temperature. The strain sensor, quantum piezo-resistive interphase sensor (sQRS), AE
sensor and DIC sensing technique has been studied by Chowdhury et al.
[217] for incremental cyclic loading. The loading were considered for
four cycles (Fig. 18) in which the first cycle of damage applied is a
classic curve indicating micro cracks [218]. From second cycle loading,
the damage curve indicates deformation due to overloading causing
damage such as micro cracking, matrix friction or fibre breakage [219].
Finally the stage of ultimate failure is shown in the linear part. In the
whole process, the measurement of strain due to damage by sQRS (Sh),
AE sensor and DIC sensing technique (∊h Armis) were well in agreement.

7.2. GPS sensors

A comparative study was performed for GPS receiver and laser

displacement meter against servo-type accelerometers [220]. It can be
observed that from Fig. 19, GPS receiver measure the same acceleration
data as that of the accelerometer thus proving a good measurement
Fig. 14. Temperature compensated strain data [53]. unit.

82
S. Das, P. Saha Measurement 129 (2018) 68–90

Fig. 18. Strain measurement for sQRS (Sh), DIC and AE sensors for cyclic load
[217].

Fig. 16. Static response evaluation of FOS and ERS [174].

Fig. 19. Acceleration measurement by GPS, laser displacement meter and ac-
celerometer [220].

Fig. 17. Practical and theoretical strain vs steel temperature for one cycle
temperature test [174].

A comparative study of DGPS and PPP has been performed on a Fig. 20. Horizontal displacement time series from DGPS and PP processing with
FFT [128].
beam under dynamic behaviour to find out the efficiency of the
methods [128]. The time series of the DGPS and PPP derived horizontal
displacements along with the Fast Fourier Transformation (FFT) has factors for data acquisition using sensors are a number of sensors to be
been shown. It can be observed from Fig. 20 that the frequency and placed to the available sensor network and the source of power for the
displacement results are almost similar stating that PPP can be used sensors to work [35]. However, the issue of sensor placement draws the
efficiently as an alternative to DGPS method. attention of different researchers. The sensors installed in the structures
are permanent. Due to the design of complex structures, the placement
of sensor is difficult. Hence the cost of data acquisition is higher [222].
8. Optimum sensor placement (OSP)
The problem of sensor accessibility is reduced with the introduction of
wireless sensors. However, their application is limited due to the pro-
Sensors are the electronic devices which are installed on the struc-
blem of energy source and time synchronization. The sensor locations
tures to obtain detailed information on the strain, stress, acceleration,
were usually determined from the past experience and knowledge of the
deformation, etc. these data obtained are utilized for the vibrational
vibration occurrence. In many health monitoring situations, there was a
measurements and structural health monitoring (SHM) [221]. The main

83
S. Das, P. Saha Measurement 129 (2018) 68–90

need for analysis of the measured data to obtain the result of un-

The number of interpolation required for complex structure is

For less damage case, the algorithm shows the need for more

The method is a new technique has so far proved to be robust

The algorithm is not suitable for modern multi axial sensors

The algorithm includes complex mathematical calculations

number of sensors. However it may be so that the current

measured parts [221]. There are two types of error: false positives,

which can detect acceleration in all the three directions.

Large number of calculations are to be solved when the
which denotes false damage to the structure, and false negatives, in

There is an increase in calculations for optimal sensor

placement of sensor could be the optimal position.

which the damage is missed [35]. Therefore optimal placement of
sensor is necessary for higher accuracy.

placement (OSP) in case of complex structure.

technique is applied on complex structure
Some of the recent sensor placement methods used are techniques

which requires sufficient amount of time.

based on Fitness Function approach [221], Bayesian Approach [35],
Monkey Algorithm [223,224], Frequency Effective Independence Ap-
proach (FEfi) [222], Distributed Wolf Algorithm [225], Fisher In-
formation Matrix (FIM) norm [226] and Iterated Improved Reduced
System (IIRS) method [227]. The advantages and disadvantages of
these methods are tabulated in Table 6.

9. Discussion
Disadvantage

The comparative study of the performances of some advanced sen-

more.

sors for SHM has been studied. The performance of the sensors sup-
plemented by advanced technology has been discussed as follows. The
Useful for obtaining the maximum performance of sensor by cost

The method is applicable for frequencies from low range to high

This method can be used to solve problems such as non linearity,

Shows its capacity to diagnose damage for the lower incomplete

acceleration or displacement data directly, hence less chance of
comparison of the sensors studied has been given in Table 5.
Developed mainly for OSP of FBG Sensors. Does not use the
Simple and efficient with less calculations and lower risk of

Uses iterative loop to pin-point the required sensor location

• One
The shuffling strategy embedded in DWA includes higher

of the basic methods for damage detection is non-contact

high dimensionality and non-differentiability at a faster

missing out any location where sensor is not present.

convergence rate, thus less calculations are required.

searching capacity and better convergence of results

method which makes use of the high-resolution camera to capture

images of cracks. These technique makes use of vision-based dis-
placement sensors are digital image correlation (DIC), pattern
matching. These images are then superimposed with known tem-
modes with the use of Jaya Algorithm

plates to determine damage pattern. The method is simple and ef-

fective as it requires no extra information about the mode shapes
and frequency. However, the environmental factors such as sur-
rounding light, the angle of elevation of the camera and hindrances
like fog can cause a problem in acquiring the damage pattern image.
missing sensor data

A similar method was performed using video-based damage detec-

tion.
• The metallic thin film strain gauge has been tested with respect to
Advantage

reduction

the theoretical value of strain obtained. The modification made is

range

the introduction of elastic shunt sheet along with the strain gauge to
measure dynamic strain. The method showed 11.5% strain over-
estimation in theoretical value. Also, complexities were faced as the
This algorithm involves the working of wolf pack i.e., searching,

optimized damage detection for validating if the sensor layout is

evaluating with the use of Bayes risk-based performance metric.

The algorithm makes use of convergence criterion to decide the

Uses Darwinian Principle of natural selection which is better in
The algorithm uses coding method which involves the steps of

A reduced order model for OSP is developed which performs

height of the shunt columns were not taken into consideration. Thus
The algorithm uses data from Frequency Response Functions

threshold redundancy by using strain mode shapes from two

The missing mode shape value is calculated with the help of

selecting the modes and exchanging the measured structural

monkey; climb process, watch-jump process and somersault

attacking and food distribution and using swarm intelligent

(FRF) to obtain dynamic information that can be used for
The algorithms works on the platform of global detection

improvement in the shunt column design was mandatory for its

process and Modal Assurance Criterion (MAC) for sensor

practical field application. Also the measurements of strain sensors

are effected due to electromagnetic interference which may causes
normal interpolation of the neighbouring points.

distortion in the final result.

algorithm to improve for efficient for solving.

• The use of piezoelectric materials in the strain sensors have been a

common choice for the researchers. These materials provide im-
munity to electromagnetic interference which the normal strain
information than the original MA.

sensors are vulnerable to. The thick film ceramic strain sensor is an
optimum for damage detection.

economic and reliable piezo-based strain sensor with good sensi-

directions simultaneously.

tivity or gauge factor. As compared to other piezo materials, ceramic

placement of sensors

placement detection.

is shock and corrosion resistant with high elasticity and can operate
Comparative study on some of the recent OSP algorithm.

at high voltages (2 kW) and high-temperature range (−20 °C to

800C). The drawback of this sensor is that it has to be calibrated for
change in atmospheric effect, which is not always feasible.
•
Method

The traditional strain sensors used were bulky and had cabled
connection. The need for lightweight sensors and wireless sensors
were in great demand for improved SHM method. PZT sensor based
Independence Approach (FEfi)

Asynchronous-Climb Monkey

System (IIRS) Method [227]

on carbon nanomaterials was tested and was found to possess good

Distributed Wolf Algorithm

Iterated Improved Reduced

Fitness Function Approach

Fisher Information Matrix

Monkey Algorithm (MA)

mechanical properties along with high sensitivity which ensured

Bayesian Approach [35]

Algorithm (AMA) [223]

damage detection. These sensors can also be for reinforcing the

Frequency Effective

(FIM) norm [226]

cement based sensors embedded in the structures.

•
OSP Algorithm

To overcome the restrictions due to cabled sensors, wireless sensing

technology collaborated with the strain sensors. The technologies
commonly used are radio frequency and MEMS. Though these sen-
[221]

[222]

[224]

[225]

sors are reliable for damage identification, the power source of the
sensors has to be recharged periodically and the receiver has to be in
Table 6

Sl No

the range of the sensor. Various solutions have been proposed for
1

self- recharging of the sensors. These include RF harvesting, solar

84
S. Das, P. Saha Measurement 129 (2018) 68–90

power harvesting, resonant vibration harvesting and piezoelectric- theoretical value. The reason for the difference may be due to
based self-powered sensor. These sensors have to be placed ac- overestimation of mass or stiffness matrix in theoretical calcula-
cording to the suitable atmospheric conditions. tions.
• The use of AE sensors, which work on the principle of detecting • An experimental study of different strain sensors, ERS, VWSG, FBG
elastic waves due to the difference in strain energy released, have and FO, were performed in the controlled condition of which VWSG
been introduced for structural diagnostics. This method is effective gave the best result. This sensor proved to be effective for short-term
for damage occurring due to crack formation. These sensors also use monitoring. On calibrating the FBG sensors, the error in reading can
the piezoelectric, MEMS technology for damage detection. These be reduced. For long-term monitoring, where temperature change is
sensors proved to be effective even for damage detection due to low not sudden, FO sensors proved to be more robust for damage
amplitude and the MEMS sensors provided higher efficiency with identification.
lower power consumption. The MEMS sensors have a problem with • It was experimentally showed that the strain measured with help of
high drift due to atmospheric thermal conditions, which can be FO sensors is same as the theoretical value for temperature range
rectified using the narrow window of sensor response. Also, struc- 30 °C to 60 °C thus making it suitable for measurement of tem-
tures under dynamic loads, such as that of the earthquake, can also perature. Also more the length of the FO sensor, more is the accurate
be determined with the help of AE sensors. measurement. Carbon nano tube strain sensor, AE sensors and DIC
• To measure the dynamic properties of a structure effectively, ac- sensing technique have been tested for strain measurement and the
celerometers have been used. This sensor provides information results were comparable.
about the change in structural parameters due to dynamic loadings, • In a comparative study for GPS sensors, it was found that GPS
wind load or earthquake load, by measuring the change in accel- sensors are capable of detecting acceleration with the same accuracy
eration. A wide range of accelerometers have been introduced for as that of accelerometers. It was also found that both DGPS and PPP
SHM; electronic cabled, wireless, GPS based, MEMS. The accuracy are good for measurement of real-time dynamic analysis.
of wireless accelerometers has been seen to reduce due to electro- • FBG sensor can be used for corrosion monitoring provided the fiber
magnetic interference. optic is properly coated. The proper coating provided to FBG sensor
• GPS based accelerometer showed the good response for detecting is PDMS coating mixed with pH sensitive hydrogel. This coating is
damage with low vibrational frequency. But these sensors faced sensitive to change in pH and can detect corrosion in the early stage.
problems due to the high sampling rate and satellite visibility. To • OSP is a major parameter which is applied for proper and efficient
avoid this problem, integrated type sensor; strain gauge, GPS re- measurement of sensor signals. Determination of optimal sensor
ceiver, and accelerometer, the arrangement has been used to in- location for dynamic measurement with good accuracy was
crease the productivity of the SHM system. The high sampling rate achieved with the help of Fitness Function approach, Bayesian ap-
can also be countered with the use of DGPS and PPP which can be proach and FEfi approach (Table 6). However these algorithms in-
used for real-time kinematic measure. clude complex calculations which might not be suitable for complex
• MEMS accelerometers, working on capacitive and piezoresistive structures. It is observed that the animal cluster based OSP algo-
principle, have proved to be effective for the wireless sensor net- rithm for efficient for sensor placement. For fast calculation, the
work. These sensors have small bandwidth and ultra-low floor noise, AMA algorithm can be used which uses an iterative method for the
thus improving the sensitivity and power consumption. The voltage result to converge. For placement of triaxial sensors, WMA algo-
sensitivity and noise immunity make them suitable for normal SHM rithm is implemented which is slow but efficient process for sensor
and seismic applications. placement, considering six degrees of freedom for assessment.
• Corrosion sensors can be embedded in the concrete for detection of • The placement of FBG sensors is carried out using FIM norm which
early stage corrosion. The PZT sensor, which uses AE waves, can be makes use of convergence criteria using strain mode shapes to find
used as surface embedded sensor for detection of corroded rebar in OSP. The chances of the algorithm to miss out sensor location are
concrete. Galvanic couple, which uses cathode anode configuration, reduced as it does not use strain of acceleration data directly. For
can be used as corrosion sensors because of their high sensitivity. low level of damage this algorithm may show false sensor placement
However these sensors are also sensitive to environmental climate location. In order to obtain OSP for low damage region, IIRS method
change. RF signals based corrosion sensor can also be used for has been introduced which uses Jaya algorithm to detect low in-
wireless corrosion monitoring. complete modes for damage diagnosis.
• Fiber Optic based sensor is the new technology sensor which pos-
sesses multiplexing quality. This sensor can be utilized for detection 10. Conclusion
of temperature, humidity, pressure, displacement, acceleration, pH
value and magnetic flux. There are three types of sensor included A number of sensors used in SHM have been studied in this paper.
under this technology; FBG sensor, multiplexed FBG sensor and FOS Most of the sensors make use of strain data for the purpose of mon-
sensor. These sensors are versatile, cost-effective and proved to be a itoring the structure. The traditional sensors used for SHM are the
good detector of damage. electronic types; strain sensors, accelerometer, and displacement
• FP, SOFO, BOTDA and BOTDR comes under the FOS. The FP and transducers. With advancement in technology, these sensors have been
SOFO systems are of a similar kind except that in SOFO system, a modified with the use different systems; fiber optics, MEMS, and PZT.
broad LED light is used for measuring the strain difference. This These technologies have improved the quality of sensing damage and
system proved effective in finding out the probable damage region proved effective in the modern world.
when compared to theoretical strain value. High-resolution photography has been used to identify damage as it
• Out of BOTDR and BOTDA, BOTDA has been adopted for practical is a good replacement of analytical methods. The method uses pre-
use. This system can detect the difference in strain released during defined damage templates to detect damage. However, this noncontact
compressive strain or increased frictional resistance, which is not damage detection technique counters problem due to environmental
detected by FBG. This system can measure both strain and tem- effects such as illumination, vision hindrance and error due to the large
perature, but care has to be taken during installation due to its angle of elevation of the camera. Also acquiring high-resolution pho-
fragile nature. tographs for small and restricted access is difficult. These sensors have
• FBG based accelerator has been used due to its ability of electro- been successful for different building structures situated at a good
magnetic immunity. This accelerometer has proved to be effective as distance while considering the light conditions to be stable and constant
it gave an average error of 10.3% when compared with the throughout.

85
S. Das, P. Saha Measurement 129 (2018) 68–90

The traditional strain sensors used are the bulky cabled electronic in structure, IIRS method which uses Jaya algorithm has been devel-
strain gauge used to measure the difference in strain occurring in the oped for OSP.
structure. These sensors are the most common type used for damage With the advances in technology, the multiplexed sensor such as
detection in any structures. Thin film strain sensors can be used for FBG sensors will be widely used in all the field of SHM. Wireless GPS
dynamic strain measurement using supporting columns. However the sensor is advancement in the field of SHM for precise damage diagnosis
bending and forces on supporting columns have to be considered for without any direct connection to power source. Sensor placement is
measurement which in real measurement is difficult. CBSS, can be used also an important aspect for efficient sensor placement such that
for column structures, bridge piers, multi storey buildings, etc. because maximum number of structural data is obtained with deployment of
of the cementitious nature which helps these sensors to bond with minimum number of sensors.
parent structure. These sensors encountered interference due to sur-
rounding environmental such as electromagnetic effect. To improve the References
efficiency of the sensor, modified sensors based on piezo and fiber
optics have been researched on. These sensors are immune to the [1] W. Hong, Z.J., G. Wu, Z. Wu, Comprehensive comparison of macro-strain mode
electromagnetic influence of the surrounding atmospheric conditions. and displacement mode based on different sensing technologies, Mechanical Syst.
Signal Process. 50–51 (2015) 563–579.
These sensors are mostly wired to a power source for constant supply of [2] L.I. Hui, O.U. Jinping, Structural Health Monitoring: From Sensing Technology
power. In this view, wireless strain sensors have been developed and Stepping to Health Diagnosis, Procedia Eng. 14 (2011) 753–760.
deployed to remotely detect damage where access for cabled sensor is [3] T.H. Yi, L.H.N., Innovative structural health monitoring technologies,
Measurement 88 (2016) 343–344.
not feasible, mainly offshore structures. [4] V. Arjun, S.B., B. Purna Chandra Rao, C.K. Mukhopadhyay, T. Jayakumar,
The piezo materials, immune to electromagnetic interference, used Optimisation of pulsed eddy current probe for detection of sub-surfacedefects in
for sensor technology are thick ceramic materials, carbon nano mate- stainless steel plates, Sens. Actuators A 226 (2015) 69–75.
[5] M. Demetgul, S.V.Y., R. Uyandik, I.N. Tansel, O. Yazicioglu, Evaluation of the
rials and MEMS. These materials have proved to be reliable for practical health of riveted joints with active and passive structural health monitoring
applications. Nano material based sensors are usually applicable for techniques, Measurement 69 (2015) 42–51.
cyclic loading owing to their grapheme epoxy structure. The AE sensor [6] Q. Han, X.J., A. Carpinteri, Localization of acoustic emission sources instructural
health monitoring of masonry bridge, Structural Control Health Monit. 22 (2015)
technology in collaboration with piezo materials and MEMS technique
314–329.
has also proved to be efficient for crack analysis in structures under [7] F. Li, P.H., G. Meng, Quantitative damage image construction in platestructures
dynamic loads. The detection of the dynamic properties of the structure using a circular PZT array and lamb waves, Sens. Actuators A 214 (2014) 66–73.
with the help of strain sensor was studied. But the prototype used [8] D. Roach, Real time crack detection using mountable comparative vacuum mon-
itoring sensors, Smart Struct. Syst. 5 (2009) 317–328.
needed improvement for practical application. [9] V.G.M. Annamdas, S.C.K., Application of electromechanical impedance technique
To measure the dynamic properties, loading due to wind pressure or for engineering structures: review and future issues, J. Intell. Mater. Syst. Struct.
earthquakes, accelerometers have been practically assigned. A wide 21 (2010) 41–59.
[10] J. Zhao, B.T., U. Amjad, Optical fiber sensing of small cracks in isotropic homo-
range of accelerometers have been introduced for SHM; electronic ca- geneous materials, Sens. Actuators A 225 (2015) 133–138.
bled, wireless, GPS based, MEMS. Out of these, GPS based and MEMS- [11] K. Sun, M.G., F. Li, L. Ye, L. Ye, Damage identification in thick steel beam based on
based accelerometers have shown accurate measurements with reduced guided ultrasonic waves, J. Intell. Mater. Syst. Struct. 21 (2010) 225–232.
[12] B.A. Sundaram, K. Ravisankar, R. Senthil, S. Parivalla, Wireless sensors for
noise interference. The MEMS-based technology for sensor develop- structural health monitoring and damage detection techniques, Curr. Sci. 104 (11)
ment has been a breakthrough in monitoring techniques. These sensors (2013) 1496–1505.
have been effective in measuring data under different noise environ- [13] C. Haynes, M. Todd, Enhanced damage localization for complex structures through
statistical modeling and sensor fusion, Mech. Syst. Sig. Process. 54 (2015)
ment and can be used for both strain and acceleration data acquisition. 195–209.
Moreover, these sensors can be also integrated with wireless tech- [14] J.D. Achenbach, Structural health monitoring—what is the prescription? Mech.
nology, giving them advantages for offshore or remote structural health Res. Commun. 36 (2009) 137–142.
[15] D.D.L. Mascarenas, F.E.B., M.D. Todd, T.G. Overly, K.M. Farinholt, G. Park,
monitoring. GPS sensor can be used for dynamic monitoring of high rise
C.R. Farrar, Development of capacitance-based and impedance-based wireless
buildings as this type of sensors are mainly dependant on satellite data sensors and sensor nodes for structural health monitoring applications, J. Sound
provided. The PPP and DGPS sensors are good for measuring damage Vib. 329 (2009) 2410–2420.
under earthquake loads and are able to work under high sampling [16] S. Das, P. Saha, S.K. Parto, Vibration based damage detction techniques used for
health monitoring of structrues: a review, J. Civil Struct. Health Monit. 6 (3)
frequency. PZT transducers and Galvanic cathode anode configuration (2016) 477–507.
can be used for detection of corrosion in steel rebars. [17] Y.Y. Li, Hypersensitivity of strain-based indicators for structural damage identi-
The last type of sensor discussed is the Fiber Optics based sensors fication: a review, Mech. Syst. Sig. Process. 24 (2010) 653–664.
[18] O.S. Salawu, Detection of structural damage through changes in frequency: a re-
used for multiple damage detection. This sensor uses light as a source of view, Eng. Struct. 19 (9) (1997) 718–723.
the impulse to detect damage and can be used for long-term health [19] S. Das, P. Saha, Structural health monitoring techniques implemented on IASC-
monitoring but has to be installed with good care as the fibers are ASCE benchmark problem: a review, J. Civil Struct. Health Monit. (2018), doi: 10.
1007/s13349-018-0292-5.
fragile. This sensor is the upcoming trend in the field of health mon- [20] E.P. Carden, F.P., Vibration based condition monitoring : a review, Struct. Control
itoring. These sensors are resistant to electromagnetic interference and Health Monit. 3 (4) (2004) 355–377.
extreme temperature conditions. The long gauge sensors can be used [21] C.H.J. Fox, The location of defects in structures: a comparison of the use of natural
frequency and mode shape data, in: The 10th International Modal Analysis
monitoring of bridges under different loads. The main parameters that
Conference, 1992. San Diego, CA.
are measured are strain and temperature. However, the application of [22] A.P. Adewuyi, Z.S. Wu, N.H.M.K. Serker, Assessment of vibration-based damage
FBG is not limited to only these two parameters. It can detect humidity, identification methods using different measurement techniques, Structural Health
Monit. 8 (6) (2009) 443–461.
acceleration, displacement, corrosion and others. Therefore keeping in
[23] J. Seo, J.W. Hu, J. Lee, Summary Review of Structural Health Monitoring
mind the careful installation of these sensors, due to their fragile nature, Applications for Highway Bridges, J. Perform. Constr. Facil (2015) p. 04015072-
the sensors are robust and can be best used for the damage analysis in 1-9.
the modern day. The sensor has been widely used for monitoring of [24] J. Seo, T.M. Czaplewski, J.H. Kimn, G. Hatfield, Integrated structural health
monitoring system and multi-regression models for determining load ratings for
building and bridge structures as the strain measured is not affected due complex steel bridges, Measurement 75 (2015) 308–319.
to change in surrounding conditions, temperature and humidity. [25] J.P. Seo, B., P. Lu, T. Wipf, J. Dahlberg, Bridge rating protocol using ambient
Apart from applications of the sensors, installation of a sensor to a trucks through structural health monitoring system, Eng. Struct. 46 (2013)
569–580.
specific point is equally important for the study. Out of the few algo- [26] B. Pharse, P. Lu, T. Wipf, L. Greimann, J. Seo, Evolution of a Bridge Damage-
rithms discussed, animal cluster algorithm has seen to be a robust Detection Algorithm, Transp. Res. Record J. Transp. Res. Board 2331 (2013)
method for deciding the crucial location of bi axial and triaxial sensors 71–80.
[27] V.G.M. Annamdas, C.K. Soh, Application of electromechanical impedance
in case of complex structure. However for low level of damage detection

86
S. Das, P. Saha Measurement 129 (2018) 68–90

technique for engineering structures: review and future issues, J. Intelligent Mater. [60] J.W. Zha, B. Zhang, R.K.Y. Li, Z.M. Dang, High-performance strain sensors based
Syst. Struct. 21 (2010) 41–59. on functionalized graphene nanoplates for damage monitoring, Compos. Sci.
[28] J.L. Rose, A baseline and vision of ultrasonic guided wave inspection potential, J. Technol. 123 (2016) 32–38.
Pressure Vessel Technol. ASME 124 (2002) 273–282. [61] S.A.A. Jabir, N.K. Gupta, Condition monitoring of the strength and stability of civil
[29] S. da Silva, C.G. Gonsalez, V. Lopes Jr., Adaptive filter feature identification for structures using thick film ceramic sensors, Measurement 46 (2013) 2223–2231.
structural health monitoring in an aeronautical panel, The Society for [62] J.W. Cho, J.S. Choi, Y.S. Yoon, Electromechanical behaviour of hybrid/glass fiber
Experimental Mechanics Series (2011). composites with tension and bending, J. Appl. Polymers 83 (2002) 2447–2453.
[30] P. Antunes, R. Travanca, H. Varum, P. Andre, Dynamic monitoring and numerical [63] T.Q. Trung, N.T. Tien, D. Kim, M. Jang, O.J. Yoon, N.E.A. Lee, Flexible reduced
modelling of communication towers with FBG based accelerometers, J. Constr. graphene oxide field-effect transistor for ultrasensitive strain sensing, Adv. Funct.
Steel Res. 74 (2012) 58–62. Mater. 24 (2014) 117–124.
[31] X. Shi, Monitoring of Reinforced Concrete Corrosion, Woodhead Publishing, [64] J. Sebastian, N. Schehl, M. Bouchard, M. Boehle, L. Li, A. Lagounov, K. Lafdi,
Elsevier, 2018. Health monitoring of structural composites with embedded carbon nanotube
[32] D.G. Lee, M. Mitrovic, A. Friedman, G.P. Carman, L. Richards, Characterization of coated glass fiber sensors, Carbon 66 (2014) 191–200.
fiber optic sensors for structural health monitoring, J. Compos. Mater. 36 (2002) [65] H.A.K. Toprakci, S.K. Kalanadhabhatla, R.J. Spontak, T.K. Ghosh, Polymer nano-
1349–1366. composites containing carbon nanofibers as soft printable sensors exhibiting
[33] F. Ansari, State-of-the-art in the applications of fiber-optic sensors to cementitious strain-reversible piezoresistivity, Adv. Funct. Mater. 23 (2013) 5536–5542.
composites, Cem. Concr. Compos. 19 (1997) 3–19. [66] W. Li, X.Z. Tang, H.B. Zhang, Z.G. Jiang, Z.Z. Yu, X.S. Du, Y.W. Mai, Simultaneous
[34] P.F.C. Antunes, H.F.T. Lima, N.J. Alberto, H. Rodrigues, P.M. Pinto, R.N. Nogueira, surface functionalization and reduction of graphene oxide with octadecylamine for
H. Varum, A.G. Costa, P.S. de Brito Andra, Optical Fiber Accelerometer System for electrically conductive polystyrene composites, Carbon 49 (2011) 4724–4730.
Structural Dynamic Monitoring, Sensors 9 (2009) 1347–1354. [67] S.A. Neild, M.S. Williams, P.D. McFadden, Development of a vibrating wire strain
[35] E.B. Flynn, M.D. Todd, A Bayesian approach to optimal sensor placement for gauge for measuring small strains in concrete beams, Strain 41 (1) (2005) 3–9.
structural health monitoring with application to active sensing, Mech. Syst. Sig. [68] H.S. Saw, J.Y.R. Liew, Assessment of current methods for the design of composite
Process. 24 (2010) 891–903. columns in buildings, J. Constr. Steel Res. 53 (2) (2005) 121–147.
[36] H.M. Schreier, J.R. Braasch, M.A. Sutton, Systematic errors in digital image cor- [69] F. Azhari, N. Banthia, Cement-based sensors with carbon fibers and carbon na-
relation caused by intensity interpolation, 39 (11) (2000) 2915–292. Soc. Photo- notubes for piezoresistive sensing, Cem. Concr. Compos. 34 (2012) 866–873.
Opt. Instrum. Eng., 39 (11) (2000) 2915–2921. [70] M.Q. Sun, R.J.Y. Liew, M.H. Zhang, W. Li, Development of cement-based strain
[37] G. Crammond, S.W. Boyd, J.M. Dulieu-Barton, Speckle pattern quality assessment sensor for health monitoring of ultra high strength concrete, Constr. Build. Mater.
for digital image correlation, Opt. Lasers Eng. 51 (2013) 1368–1378. 65 (2014) 630–637.
[38] L.P. Yu, B. Pan, The errors in digital image correlation due to overmatched shape [71] F. Azhari, Cement-based sensors for structural health monitoring, in: Civil
functions, Measurement Sci. Technol. 26(4) 045202. Engineering, 2009, University of British Columbia.
[39] G. Busca, A. Cigada, P. Mazzoleni, E. Zappa, Vibration monitoring of multiple [72] P.W. Chen, D.D.L. Chung, Carbon fiber reinforced concrete for smart structures
bridge points by means of a unique vision-based measuring system, Exp. Mech. 54 capable of non-destructive flaw detection, Smart Mater. Struct. 2 (1) (1993)
(2014) 255–271. 22–30.
[40] X.W. Ye, Y.Q. Ni, T.T. Wai, K.Y. Wong, X.M. Zhang, F. Xu, A vision-based system [73] G.Y. Li, P.M. Wang, X. Zhao, Pressure-sensitive properties and microstructure of
for dynamic displacement measurement of long-span bridges: algorithm and ver- carbon nanotube reinforced cement composites, Cem. Concr. Compos. 29 (5)
ification, Smart Struct. Syst. 12 (2013) 363–379. (2007) 377–382.
[41] D. Ribeiro, R. Calcada, J. Ferreira, T. Martins, Non-contact measurement of the [74] E.G. Straser, A.S. Kiremidjian, A Modular, Wireless Damage Monitoring System for
dynamic displacement of railway bridges using an advanced video-based system, Structures, in: John A. Blume Earthquake Eng. Ctr. 1998, Stanford University:
Eng. Struct. 75 (2014) 164–180. Stanford, CA.
[42] X.W. Ye, T.H. Yi, C.Z. Dong, T. Liu, Vision-based structural displacement mea- [75] B. Graybeal, B.M. Phares, D. Rolander, M. Moore, G. Washer, Visual inspection of
surement: system performance evaluation and influence factor analysis, highway bridges, J. Nondestr. Eval. 21 (3) (2002) 67–83.
Measurement 88 (2016) 372–384. [76] Y. Zhang, L. Bai, Rapid structural condition assessment using radio frequency
[43] D. Feng, M.Q. Feng, Experimental validation of cost-effective vision-based struc- identification (RFID) based wireless strain sensor, Autom. Constr. 54 (2015) 1–11.
tural health monitoring, Mech. Syst. Sig. Process. 88 (2017) 199–211. [77] A. Wong, F. Wong, A. Nadeem, Government roles in implementing building in-
[44] R.C. Gonzalez, R.E. Woods (Eds.), Digital Image Processing, ed. 3rd. 2008, Upper formation modeling systems: comparison between Hong Kong and the United
Saddle River, NJ: Pearson Prentice Hall. States, Constr. Innovation Information Process Manage. 11 (1) (2011) 61–76.
[45] X.W. Ye, T.H. Yi, C.Z. Dong, H. Bai, Multi-point displacement monitoring of [78] U. Isikdag, G. Aouad, J., Underwood, S., Wu, Building information models: a re-
bridges using a vision based approach, Wind Struct. 20 (2) (2015) 315–326. view on storage and exchange mechanisms, in: Bringing ITC Knowledge to Work,
[46] J.G. Chen, A. Davis, N. Wadhwa, F. Durand, W.T. Freeman, O. Büyüköztürk, Video 24th W78 Conference, Maribor, 2007.
Camera-Based Vibration Measurement for Civil Infrastructure Applications, J. [79] A. Daliri, A. Galehdar, S. John, C. Wang, W. Towe, K. Ghorbani, Wireless strain
Infrastruct. Syst. B4016013 (2017). measurement using circular microstrip patch antennas, Sens. Actuators A 184
[47] J.G. Chen, N. Wadhwa, Y.J. Cha, F. Durand, W.T. Freeman, O. Buyukozturk, (2012) 86–92.
Modal identification of simple structures with high-speed video using motion [80] J.P. Lynch, Y. Wang, K.J. Loh, J., Yi, C.B. Yun, Wireless structural monitoring of
magnification, J. Sound Vib. 345 (2015) 58–71. the geumdang bridge using resolution enhancing signal conditioning, in: 24th
[48] V. Giurgiutiu, A. Cuc, Embedded non-destructive evaluation for structural health International Modal Analysis Conference (IMACXXIV), 2006, St.Louis, MO.
monitoring, damage detection, and failure prevention, Shock Vibration Digest 37 [81] T.C. Huynh, J.H. Park, J.T. Kim, Structural identification of cable-stayed bridge
(2) (2005) 83–105. under back-to-back typhoons by wireless vibration monitoring, Measurement 88
[49] J.P. Lynch, K.J. Loh, A summary review of wireless sensors and sensor networks (2016) 385–401.
for structural health monitoring, Shock Vibration Digest 38 (2) (2006) 91–128. [82] J.T. Kim, T.C. Huynh, S.Y. Lee, Wireless structural health monitoring of stay cables
[50] K. Kesavan, K. Ravisankar, S. Parivallal, P. Sreeshylam, S. Sridhar, Experimental under two consecutive typhoons, Struct. Monit. Maintenance 1 (1) (2014) 47–67.
studies on fiber optic sensors embedded in concrete, Measurement 43 (2) (2010) [83] S.Y. Lee, T.C. Huynh, J.T. Kim, Structural identification of gravity-type caisson
157–163. structure via vibration feature analysis, Smart Struct. Syst. 15 (2) (2015) 259–281.
[51] S. Yehia, T. Landolsi, M. Hassan, M. Hallal, Monitoring of strain induced by heat of [84] J.H. Yi, C.B. Yun, Comparative study on modal identification methods using
hydration, cyclic and dynamic loads in concrete structures using fiber-optics output-only information, Struct. Eng. Mech. 17 (3–4) (2004) 445–446.
sensors, Measurement 52 (2014) 33–46. [85] J.J. McCullagh, T. Galchev, R.L. Peterson, R. Gordenker, Y. Zhang, J. Lynch,
[52] J.R. Casas, P.J. Cruz, Fiber optic sensors for bridge monitoring, J. Bridge Eng. 8 (6) K. Najafi, Long-term testing of a vibration harvesting system for the structural
(2003) 362–373. health monitoring of bridges, Sens. Actuators A 217 (2014) 139–150.
[53] Y. Ge, M.Z.E.B. Elshafie, S. Dirar, C.R. Middleton, The response of embedded strain [86] K.M. Farinholt, N. Miller, W. Sifuentes, J. MacDonald, C.R. Farrar, Energy har-
sensors in concrete beams subjected to thermal loading, Constr. Build. Mater. 70 vesting and wireless energy transmission for embedded SHM sensor nodes, Struct.
(2014) 279–290. Health Monit. 9 (2010) 269–280.
[54] N.A. Hoult, P.J. Bennett, C.R. Middleton, K. Soga, Distributed fibre optic strain [87] T. Galchev, J. McCullagh, R.L. Peterson, K. Najafi, A. Mortazawi, Energy har-
measurements for pervasive monitoring of civil infrastructure, in: International vesting of radio frequency and vibration energy to enable wireless sensor mon-
conference on structural health monitoring of intelligent infrastructure (SHMII-4), itoring of civil infrastructure, in: SPIE 7983, Nondestructive Characterization for
H.B. Meier Urs, Motavalli Masoud, Editor. 2009: Zurich, Switzerland. Composite Materials, Aerospace Engineering, Civil Infrastructure, and Homeland
[55] Q. Xia, F. Quail, Principles and validation of strain gauge shunt design for large- Security 2011, 798314 2011.
dynamic strain measurement, Sens. Actuators A 241 (2016) 124–134. [88] E. Sazonov, L. Haodong, D. Curry, P. Pilay, Self-powered sensors for monitoringof
[56] L.S. Katafygiotis, K.V. Yuen, Bayesian spectral density approach for modal up- highway bridges, IEEE Sensors 9 (2009) 1422–1429.
dating using ambient data, Earthquake Eng. Struct. Dyn. 30 (8) (2001) [89] S.D. Kwon, J. Park, K. Law, Electromagnetic energy harvester with re-
1103–1123. pulsivelystacked multilayer magnets for low frequency vibrations, Smart Mater.
[57] F. Tegtmeier, M. Peters, Determination of structural damage to buildings using Struct. 22 (2013) 055007.
microsensors based on strain gauge technology, in: The XVI IMEKO World [90] M. Peigney, D. Siegert, Piezoelectric energy harvesting from traffic-induced bridge
Congress, 2000, Vienna. vibrations, Smart Mater. Struct. 22 (2013) 095019.
[58] T. Wada, S.J. Stein, M.A. Stein, S.A. Chitale, The state-of-the-art of thick film [91] N. Elvin, A. Elvin, D.H. Choi, A self-powered damage detection sensor, J. Strain
technology for automotive sensors, IEMT/IMC (1997). Anal. Eng. Des. 38 (2) (2003) 115–124.
[59] B. Morten, M. Prudenziati, Piezoresistive thick-film sensors, Handbook of Sensors [92] N. Lajnef, S. Chakrabartty, N. Elvin, A piezo-powered floating-gate sensor array for
and Actuators, 1994. long-term fatigue monitoring in biomechanical implants, IEEE Trans. Biomed.

87
S. Das, P. Saha Measurement 129 (2018) 68–90

Circuits Syst. 2 (3) (2008) 164–172. Monitoring and Evaluation Using GPS-Time Series Monitoring Systems: A Review,
[93] M. Rahimi, H. Shah, G.S. Sukhatme, J. Heideman, D. Estrin, Studying the feasi- Int. J. Geo-Information (2017) 6(382).
bility of energy harvesting in a mobile sensor network, in: 2003 IEEE International [127] J.F. Zumberge, M.B. Heflin, D.C. Jefferson, M.M. Watkins, F.H. Webb, Precise
Conference on Robotics & Automation, 2003, Taipei, Taiwan. point positioning for the efficient and robust analysis of GPS data from large
[94] K. Ono, Application of acoustic emission for structure diagnosis, Diagnost.- networks, J. Geophys. Res. Solid Earth 102 (1997) 5005–5017.
Diagnost. Struct. Health Monit 2 (58) (2011) 3–18. [128] C.O. Yigit, E. Gurlek, Experimental testing of high-rate GNSS precise point posi-
[95] A. Nair, C. Cai, Acoustic emission monitoring of bridges: review and casestudies, tioning (PPP) method for detecting dynamic vertical displacement response of
Eng. Struct. 32 (2010) 1704–1714. engineering structures, Geomatics Nat. Hazards Risk (2017) 1–12.
[96] T. Schumacher, C. Higgins, S. Lovejoy, Acoustic emission monitoring ofconven- [129] H. Han, J. Wang, X. Meng, H. Liu, Analysis of the dynamic response of a long span
tionally reinforced concrete highway bridges under service conditions, bridge using GPS/accelerometer/anemometer under typhoon loading, Eng. Struct.
Nondestructive Testing Mater. Struct. Springer 6 (2013) 847–853. 122 (2016) 238–250.
[97] H. Saboonchi, D. Ozevin, M. Kabir, MEMS sensor fusion: acoustic emission and [130] V.K. Ingle, J.G. Proakis, Digital signal processing using Matlab, PWS Publishing
strain, Sens. Actuators A 247 (2016) 566–578. Company, Boston (US), 1997.
[98] A. Gallego, E. Martínez, Emisión Acústica. Niveles I y II AEND. 2016: FC Editorial. [131] G.W. Roberts, E. Cosser, X. Meng, A. Dodson, High Frequency Deflection
[99] V.F. Godínez-Azcuaga, O. Ley, Structural health monitoring of aerospace struc- Monitoring of Bridges by GPS, J. Global Positioning Syst. 3 (1–2) (2004) 226–231.
tures using a low-power, wireless multichannel acoustic emission node, in: 5th [132] A. Lepadatu, C. Tiberius, GPS for structural health monitoring – case study on the
International Symposium on NDT in Aerospace, 2013, Singapore. Basarab overpass cable-stayed bridge, J. Appl. Geodesy 8 (1) (2014) 65–85.
[100] F. Sagasta, M.E. Zitto, R. Piotrkowski, A. Benavent-Climent, E. Suarez, A. Gallego, [133] J. Bogusz, M. Figurski, G. Nykiel, M. Szolucha, M. Wrona, GNSS-based multi-
Acoustic emission energy b-value for local damage evaluation in reinforced con- sensor system for structural monitoring applications, J. Appl. Geodesy 6 (2012)
crete structures subjected to seismic loadings, Mech. Syst. Sig. Process. 102 (2017) 55–64.
262–277. [134] S. Kavitha, R.J. Daniel, K. Sumangala, A simple analytical design approach based
[101] O. Yapar, P.K. Basu, P. Volgyesi, A. Ledeczi, Structural health monitoring of on computer aided analysis of bulk micromachined piezoresistive MEMS accel-
bridges with piezoelectric AE sensors, Eng. Fail. Anal. 56 (2015) 150–169. erometer for concrete SHM applications, Measurement 46 (2013) 3372–3388.
[102] P. Moore, R. Miller, R.K. Hill, Acoustic emission testing. Nondestructive testing [135] J.P. Lynch, A. Patridge, K.H. Law, W. Thomas, W. Kenny, A.S. Kiremidjian,
handbook, ed. A.S.f.N. Testing, Vol. 6, 2005. E. Carryer, Design of piezoresistive MEMS based accelerometer for integration
[103] A.W. Harris, I.J. Oppenheim, D.W. Greve, MEMS-based high-frequency vibration with wireless sensing unit for structural monitoring, J. Aerosp. Eng. 16 (2003)
sensors, Smart Mater. Struct. 20 (7) (2011). 108–114.
[104] H. Saboonchi, D. Ozevin, MEMS acoustic emission transducers designed with high [136] S. Kavitha, R.J. Daniel, K. Sumangala, High performance MEMS accelerometers for
aspect ratio geometry, Smart Mater. Struct. 22 (2013). concrete SHM applications and comparison with COTS accelerometers, Mech.
[105] H. Saboonchi, D. Ozevin, Metal MUMPs-based piezoresistive strain sensors for- Syst. Sig. Process. 66–67 (2015) 410–424.
integrated on-chip sensor fusion, Sensors 15 (2015) 568–578. [137] S. Kavitha, R.J. Daniel, K. Sumangala, Design and Analysis of MEMS Comb Drive
[106] A.E. Aktan, F.N. Catbas, K.A. Grimmelsman, M. Pervizpour, Develpoment of a Capacitive Accelerometer for SHM and Seismic Applications, Measurement 93
Model Health Monitoring Guide for Major Bridges, in: Federal Highway (2016) 327–339.
Administration Research and Development, 2003, Drexel University. [138] R.G. Walmsley, L.K. Kiyarna, D.M. Milligan, R.L. Alley, D.L. Erickson,
[107] Y. Bao, J.L. Beck, H. Li, Compressive sampling for accelerometer signals in P.G. Hartwell, Micro-G silicon accelerometer using surface electrodes, IEEE -
structural health monitoring, Struct. Health Monit. 10 (3) (2011) 235–246. SENSORS (2009) 971–974.
[108] W.U. Bajwa, J. Haupt, A.M. Sayeed, R. Nowak, Compressive wireless sensing, in: [139] M.H. Bao, S. Middelhoek, Handbook of Sensors and Actuators, Micromechanical
Proceedings of the 5th International Conference on Information Processing in Transducers Vol. 8 Elsevier, 2000.
Sensor Networks (IPSN’06), 2006, Nashville, Tamil Nadu. [140] A. Vogl, D.T. Wanga, P. Storasa, T. Bakkea, M.M.V. Taklo, A. Thomson, L. Balgard,
[109] D. Takhar, J.N. Laska, M.B. Wakin, M.F. Duarte, D. Baron, S. Sarvotham, F.K. Design, process and characterisation of a high-performance vibration sensor for
Kelly, R.G. Baraniuk, A new compressive imaging camera architecture using op- wireless condition monitoring, Sens. Actuators A 153 (2) (2009) 155–161.
tical-domain compression, in: Proceedings of the Computational Imaging IV at [141] C. Acar, A.M. Shkel, Experimental evaluation and comparative analysis of com-
SPIE Electronic Imaging, 2006, San Jose, California. mercial variable - capacitance MEMS accelerometers, J. Micromech. Microeng. 13
[110] S. Kirolos, J. Laska, M. Wakin, M. Duarte, D. Baron, T. Ragheb, Y. Massoud, R. (2003) 634–645.
Baraniuk, Analog-to-information conversion via random demodulation, in: [142] C. Acar, A.M. Shkel, Comparative Characterization of Low-g Capacitive MEMS
Proceedings of the IEEE Dallas Circuits and Systems Workshop, 2006, Dallas, Accelerometers. Structural Health Monitoring from Diagnostics and Prognostics to
Texas. Structural Health Management. 2003, USA: D Estech Publications.
[111] M.J. Chae, H.S. Yoo, J.Y. Kim, M.Y. Cho, Development of a wireless sensor net- [143] ADXL 150/250 Specification Data Sheets. Analog Devices Inc, 1998.
work system for suspension bridge health monitoring, Autom. Constr. 21 (2012) [144] E.I. Gonenil, Z.C. Butler, P. Donald, Surface micromachined MEMS accelerometer
237–252. on flexible polymide substrate, IEEE Sensors 11 (2011) 2318–2326.
[112] G. Kilic, M.S. Unluturk, Testing of wind turbine towers using wireless sensor [145] S. Kavitha, R.J. Daniel, K. Sumangala, Design and analysis of bulk micromachined
network and accelerometer, Renew. Energy 75 (2015) 318–325. piezoresistive MEMS accelerometer for concrete SHM applications, Digital Sensors
[113] L. Fushun, L. Huajun, L. Wei, W. Bin, Experimental study of improved modal strain Sensor Syst. Practical Des. 144 (9) (2012) 62–75.
energy method for damage localisation in jacket-type offshore wind turbines, [146] G. Li, A.A. Tseng, Transient and impact dynamics of a Micro-accelerometer, J.
Renew. Energy 72 (2014) 174–181. Mater. Process. Manuf. Sci. 9 (2000) 143–156.
[114] D. Zhu, Y. Wang, Brownjohn, Vibration testing of a steel girder bridge using cabled [147] L. Abosrra, A.F. Ashour, M. Youseffi, Corrosion of steel reinforcement in concrete
and wireless sensors, Front. Architecture Civil Eng. China 5 (249) (2011). of different compressive strengths, Constr. Build. Mater. 25 (2011) 3915–3925.
[115] S.B. Im, S. Hurlebaus, Y.J. Kang, Summary Review of GPS Technology for [148] Y. Cheng, A. Hanif, E. Chen, G. Ma, Z. Li, Simulation of a novel capacitive sensor
Structural Health Monitoring, J. Struct. Eng. 141 (2013) 1653–1664. for rebar corrosion detection, Constr. Build. Mater. 174 (2018) 613–624.
[116] K.Y. Wong, K.L. Man, W.Y. Chan, Monitoring Hong Kong’s bridges: Realtime ki- [149] K.Y. Ann, H.W. Song, Chloride threshold level for corrosion of steel in concrete,
nematic spans the gap, GPS World 7 (2001) 10–18. Corros. Sci. 49 (11) (2007) 4113–4133.
[117] X. Meng, G.W. Roberts, A.H. Dodson, C.J. Brown, GNSS for bridge deformation: [150] X. Shi, N. Xie, K. Fortune, J. Gong, Durability of steel reinforced concrete in
Limitations and solutions, in: Third International Conference on Bridge chloride environments: an overview, Constr. Build. Mater. 30 (2012) 125–138.
Maintenance, Safety and Management, 2006, Porto, Portugal. [151] J.A. González, A. Molina, M.L. Escudero, C. Andrade, Errors in the electrochemical
[118] X. Meng, A.H. Dodson, G.W. Roberts, Detecting bridge dynamics with GPS and evaluation of very small corrosion rates-I. Polarization resistance method applied
triaxial accelerometers, Eng. Struct. 29 (2007) 3178–3184. to corrosion of steel in concrete, Corrosion Sci. 25 (1985) 917–930.
[119] F. Moschas, S. Stiros, Noise characteristics of high-frequency, short-duration GPS [152] C. Liu, Q. Bi, A. Leyland, A. Natthews, An electrochemical impedance spectroscopy
records from analysis of identical, collocated instruments, Measurement 46 (2013) study of the corrosion behaviour of PVD coated steels in 0.5 N NaCl aqueous so-
1488–1506. lution: Part II.: EIS interpretation of corrosion behaviour, Corros. Sci. 45 (2003)
[120] T.H. Yi, H.N. Li, M. Gu, Experimental assessment of high-rate GPS receivers for 1257–1273.
deformation monitoring of bridge, Measurement 46 (2013) 420–432. [153] S. Sathiyanarayanan, P. Natarajan, K. Saravanan, S. Srinivasan, Corrosion mon-
[121] T.H. Yi, H.N. Li, M. Gu, Full scale measurement of dynamic response of a sus- itoring of steel in concrete by galvanostatic pulse technique, Cem. Concr. Compos.
pension bridge subjected to environmental loads using GPS technology, Sci, China 28 (2006) 630–637.
Technol. Sci. 53 (2) (2010) 469–479. [154] W. Li, Q. Kong, S.C.M. Ho, Y.L. Mo, G. Song, Feasibility study of using smart
[122] T.H. Yi, H.N. Li, M. Gu, Characterization and extraction of global positioning aggregates as embedded acoustic emission sensors for health monitoring of con-
system multipath signals using an improved particle-filtering algorithm, Meas. Sci. crete structures, Smart Mater. Struct. 25 (11) (2016) 115031.
Technol. 22 (075101) (2011) p. 11 pp. [155] L. Qin, H. Ren, B. Dong, F. Xing, Development of technique capable of identifying
[123] M.R. Kaloop, J.W. Hu, Dynamic Performance Analysis of the Towers of a Long- different corrosion stages in reinforced concrete, Appl. Acoust. 94 (2015) 53–56.
Span Bridge Based on GPS Monitoring Technique, J. Sensors 2016 (2016) 14 [156] V. Talakokula, S. Bhalla, A. Gupta, Corrosion assessment of reinforced concrete
pages. structures based on equivalent structural parameters using electro-mechanical
[124] F. Moschas, S. Stiros, Dynamic deflections of a stiff footbridge using 100-Hz GNSS impedance technique, J. Intell. Mater. Syst. Struct. 25 (4) (2014) 484–500.
and accelerometer data, J. Surv. Eng. 141 (2015) 04015003. [157] J.H. Yoo, Z.T. Park, J.G. Kim, L. Chung, Development of a galvanic sensor system
[125] F. Moschas, A. Avallone, V. Saltogianni, S. Stiros, Strong motion displacement for detecting the corrosion damage of the steel embedded in concrete structures:
waveforms using 10-Hz precise point positioning GPS: An assessment based on part 1. Laboratory tests to correlate galvanic current with actual damage, Cem.
free oscillation experiment, Int. Assoc. Earhquake Eng. 44 (2015) 657–675. Concr. Compos. 33 (12) (2003) 2057–2062.
[126] M.R. Kaloop, E. Elbeltagi, J.W. Hu, A. Elrefai, Recent Advances of Structures [158] E.V. Pereira, R.B. Figueira, M.M.L. Salta, I.T.E. Da Fonseca, A galvanic sensor for

88
S. Das, P. Saha Measurement 129 (2018) 68–90

monitoring the corrosion condition of the concrete reinforcing steel: relationship [189] Y.Q. Ni, Y.X. Xia, Strain-Based Condition Assessment of a Suspension Bridge
between the galvanic and the corrosion currents, Sensors 9 (11) (2009) Instrumented with Structural Health Monitoring System, Int. J. Struct. Stability
8391–8398. Dyn. 16 (2016)(1640027) 23.
[159] Z.T. Park, Y.S. Choi, J.G. Kim, L. Chung, Development of a galvanic sensor system [190] C.K. Lau, K.Y. Wong, K.S. Ho, The wind and structural health monitoring system
for detecting the corrosion damage of the steel embedded in concrete structure: (WASHMS) for the three longspan cable-supported bridges in Hong-Kong. IABSE
Part 2. Laboratory electrochemical testing of sensors in concrete, Cem. Concr. Res. congress report = Rapport du congrès, AIPC = IVBH Kongressbericht 15 (1996)
35 (9) (2005) 1814–1819. 104–112.
[160] K. Perveen, G.E. Bridges, S. Bhadra, D.J. Thomson, An embedded inductively [191] T.H.T. Chan, L. Yu, H.Y. Tam, Y.Q. Ni, S.Y. Liu, W.H. Chung, L.K. Cheng, Fiber
coupled printed circuit board based corrosion potential sensor, in: Instrumentation Bragg grating sensors for structural health monitoring of Tsing Ma bridge: back-
and Measurement Technology Conference (I2MTC), 2013 IEEE International, ground and experimental observation, Eng. Struct. 28 (2006) 648–659.
IEEE, 2013, Piscataway, NJ. [192] C. Rodrigues, F.C. Lage A, J. Figueiras, Development of a long-term monitoring
[161] S. Bhadra, D.J. Thomson, G.E. Bridges, A wireless embedded passive sensor for system based on FBG sensors applied to concrete bridges, Eng. Struct. 32 (8)
monitoring the corrosion potential of reinforcing steel, Smart Mater. Struct. 22 (7) (2010) 1993–2002.
(2013) 075019. [193] G.D. Zhou, T.H. Yi, B. Chen, Innovative Design of a Health Monitoring System and
[162] J.J. Myers, E. Hernandez, Implementation of radio frequency identification (RFID) Its Implementation in a Complicated Long-Span Arch Bridge, J. Aerosp. Eng.
sensors for monitoring of bridge deck corrosion in Missouri., in Final Report for B4016006 (2016) p. 17 pp.
the Center for Transportation Infrastructure and Safety/NUTC Program. 2014, [194] B. Wu, G. Wu, H. Lu, D. Feng, Stiffness monitoring and damage assessment of
Missouri University of Science and Technology: Rolla, MO. bridges under moving vehicular loads using spatially-distributed optical fiber
[163] S.G.N. Murthy, Batteryless Wireless RFID based embedded sensors for long term sensors, Smart Mater. Struct. 26 (2017)(035058) (2017) p. 21 pp.
monitoring of reinforced concrete structures. in International Symposium Non- [195] T. Ojio, K. Yamada, Bridge weigh-in-motion systems using stringers of plate girder
Destructive Testing, Civ. Eng. (2015). bridges, in: 3rd Int. Conf. on Weigh-In- Motion (ICWIM3), 2002.
[164] G. Uva, F. Porco, A. Fiore, G. Porco, Structural monitoring using fiber optic sensors [196] C. Connolly, Fiber-optic-based sensors bring new capabilities to structural mon-
of apre-stressed concrete viaduct during construction phases, Case Stud. itoring, Sensor Rev. 26 (3) (2006) 236–243.
Nondestr.Test. Eval. 2 (2014) 27–37. [197] E. Shehata, S. Rizkalla, Intelligent sensing for innovative bridges, J. Intell. Mater.
[165] E. Al-Fakih, The use of fiber Bragg grating sensors in biomechanics and re- Syst. Struct. 10 (4) (1999) 304–313.
habilitation applications: the state-of-the-art and ongoing research topics, Sensors [198] A. Elamari, D. Inaudi, L. Pflug, N. Gisin, J. Breguet, S. Vurpillot, Low-coherence
(2012) 12. deformation sensors for the monitoring of civil engineering structures, Sensors
[166] B. Torres, I.P. Zaforteza, P.A. Calderón, J.M. Adam, Analysis of the strain transfer Actuators A Phys. 44 (1994) 125–130.
in a new FBG sensor for Structural Health Monitoring, Eng. Struct. 33 (2011) [199] D. Inaudi, N. Casanova, S. Vurpillot, P. Kronenberg, G. Martinola, G. Steinmann, J.
539–548. F. Mathier, SOFO: Structural Monitoring with Fiber Optic Sensor, in: Monitoring
[167] L.K. Cheng, P.M. Toet, Innovative fiber bragg grating sensors for highly de- and Safety Evaluation of Existing Concrete Structures, 1999, Vienna, Austria.
manding applications: Considerations, concepts, and designs, Opto-Mechanical [200] D.Z. Stupar, J.S. Bajic, B.M. Dakic, M.P. Slankamenac, M.B. Zˇivanov, The possi-
Fiber Optic Sensors, 2018: Elsevier. bility of using a plastic optical fibre as a sensing element in civil structural health
[168] H.N. Li, D.S. Li, G.B. Song, Recent applications of fiber optic sensors to health monitoring, Phys. Scr. 2013 (2013) 014031.
monitoring in civil engineering, Eng. Struct. 26 (11) (2004) 1647–1657. [201] C.Y. Hong, Y.F. Zhang, L.Q. Liu, Application of distributed optical fiber sensor for
[169] S. Delepine-Lesoille, E. Merliot, C. Boulay, L. Quétel, M. Delaveau, A. Courteville, monitoring the mechanical performance of a driven pile, Measurement 88 (2016)
Quasi-distributed optical fibre extensometers for continuous embedding into 186–193.
concrete: design and realization, Smart Mater. Struct. 15 (4) (2006) 931. [202] R.M. Measures, Structural monitoring with fibre optic technology, Elsevier, Editor,
[170] P. Bourne-Webb, B. Amatya, K. Soga, T. Amis, C. Davidson, P. Payne, Energy pile 2001, Academic Press, p. 716.
test at Lambeth College, London: geotechnical and thermodynamic aspects of pile [203] A. Misra, C.H. Chen, Analytical solution for micropile design under tension and
response to heat cycles, Geotechnique 59 (2009) 237–248. compression, Geotech. Geol. Eng. 22 (2004) 199–225.
[171] M. Majumder, T.M. Gangopadhyay, A.K. Chakraborty, Fibre Bragg gratings in [204] C.K.Y. Leung, K.T. Wan, L. Chen, A novel optical fiber sensor for steel corrosion in
structural health monitoring—present status and applications, Sens. Actuators, A concrete structures, Sensors 8 (2008) 1960–1976.
147 (1) (2008) 150–164. [205] T.H. Nguyen, T. Venugopala, S. Chen, T. Sun, K.T. Grattan, S.E. Taylor,
[172] C. Rodrigues, C.F. Félix, J. Figueiras, FBG based strain monitoring in the re- P.A.M. Basheer, A.E. Long, Fluorescence based fibre optic pH sensor for the pH
habilitation of a centenary metallic bridge, Eng. Struct. 44 (2012) 281–290. 10–13 range suitable for corrosion monitoring in concrete structures, Sens.
[173] F. Ansari, Practical implementation of optical fiber sensors in civil structural Actuators B 191 (2014) 498–507.
health monitoring, J. Intell. Mater. Syst. Struct. 18 (8) (2007) 879–889. [206] M. Torres-Luque, E. Bastidas-Arteaga, F. Schoefs, M. Sánchez-Silva, J.F. Osma,
[174] B.J.A. Costa, J.A. Figueiras, Fiber optic based monitoring system applied to a Non-destructive methods for measuring chloride ingress into concrete: state-of-
centenary metallic arch bridge: design and installation, Eng. Struct. 44 (2012) the-art and future challenges, Constr. Build. Mater. 68 (2014) 68–81.
271–280. [207] C.H. Tan, Y.G. Shee, B.K. Yap, F.R.M. Adikan, Fiber Bragg Grating Based Sensing
[175] P. Moyo, J.M.W. Brownjohn, R. Suresh, S.C. Tjin, Development of fiber Bragg System: Early Corrosion Detection for Structural Health Monitoring, Sens.
grating sensors for monitoring civil infrastructure, Eng. Struct. 27 (12) (2005) Actuators A 246 (2016) 123–128.
1828–1834. [208] A.D. Kersey, M.A. Davis, H.J. Patrick, M. LeBlanc, K.P. Koo, C.G. Askins,
[176] H. Adeli, X. Jiang, Dynamic fuzzy wavelet neural network model for structural M.A. Putnam, E.J. Friebele, Fibre grating sensors, J. Lightwave Technol. 15 (1997)
system identification, J. Struct. Eng. 132 (1) (2006) 102–111. 1442–1463.
[177] B.K.A. Ngoi, J. Paul, L.P. Zhao, Z.P. Fang, Enhanced lateral pressure tuning of fiber [209] I.D. Johnson, D.K. McCluskey, C.K.L. Tan, M.C. Tracey, Mechanical
Bragg gratings by polymer packaging, Opt. Commun. 242 (2004) 425–430. Characterization of Bulk Sylgard 184 for Microfluidics and Microengineering, J.
[178] T.A. Dawood, R.A. Shenoi, M. Sahin, A procedure to embed fibre Bragg grating Micromech. Microeng. 24 (2014) 1–7.
strain sensors into GFRP sandwich structures, Composites A 38 (2007) 217–226. [210] C.H. Tan, F.R.M. Adikan, Y.G. Shee, B.K. Yap, Non-destructive fiber Bragg grating
[179] W. Chung, D. Kang, Full-scale test of a concrete box girder using FBG sensing based sensing system: early corrosion detection for structural health monitoring,
system, Eng. Struct. 30 (2008) 643–652. Sens. Actuators A 268 (2017) 61–67.
[180] J.E. Dally, W.F. Riley, Experimental stress analysis, 2nd Edition., McGraw-Hill, [211] T. Sharma, G. Madras, Effect of crosslinker on the swelling and adsorption prop-
1978. erties of cationic superabsorbent, Indian Acad. Sci. 39 (2016) 613–626.
[181] G. Kister, D. Winter, R.A. Badcock, Y.M. Gebremichael, W.J.O. Boyle, B.T. Meggitt, [212] T. Rytter, Vibration Based Inspection of Civil Engineering Structure, in
K.T.V. Grattan, G.F. Fernando, Structural health monitoring of a composite bridge Department of Building Technology and Structure Engineering, Aalborg
using Bragg grating sensors. Part 1: evaluation of adhesives and protection systems University, Denmark, 1993.
for the optical sensors, Eng. Struct. 29 (2007) 440–448. [213] T. Kleckers, B. Günther, Optical Versus Electrical Strain Gages: A Comparison,
[182] H.K. Kang, J.H. Bang, C.S. Hong, C.G. Kim, Simultaneous measurement of strain, HBM Measurements (2008).
temperature and vibration frequency using a fiber optic sensor, J. Measurement [214] J. Roths, A. Wilfert, P. Kratzer, F. Jülich, R. Kuttler, Strain calibration of
Sci. Technol. 13 (2002) 1191–1196. opticalFBG-based strain sensors. (EWOFS’10) Fourth European workshop on op-
[183] C. Fernández-Valdivielso, I.R. Matías, F.J. Arregui, Simultaneous measurement of tical fibre sensors, Int. Soc. Optics Photonics (2010) 76530F-F-4.
strain and temperature using a fiber Bragg grating and a thermochromic material, [215] P. Sreeshyiarn, K. Ravisankar, S. Parivallal, K. Kesavan, S. Sridhar, Condition
Sens. Actuators A 101 (2002) 107–116. monitoring of prestressed concrete structures using vibrating wire sensors, Int. J.
[184] T. Poloso, Fibre Bragg optical sensing technology, Smart Mater. (2001) 7–10. COMADEM 11 (3) (2008) 46–54.
[185] D. Li, Z. Zhou, J. Ou, Dynamic behavior monitoring and damage evaluation for [216] A.M. Neville, Properties of concrete, ed. P. Hall, 1997.
arch bridge suspender using GFRP optical fiber Bragg grating sensors, Opt. Laser [217] S.N. Chowdhury, H. Bellegou, M. Castro, P. Longrais, J.F. Feller, Crossed in-
Technol. 44 (2012) 1031–1038. vestigation of damage in composites with embedded quantum resistive strain
[186] D.S. Li, Z. Zhou, J.P. Ou, Development and sensing properties study of FRP–FBG sensors (sQRS), acoustic emission (AE) and digital image correlation (DIC),
smart stay cable for bridge health monitoring applications, Measurement 44 Compos. Sci. Technol. 160 (2018) 79–85.
(2011) 722–729. [218] S.N. Chowdhury, H. Bellegou, I. Pillin, M. Castro, P. Longrais, J.F. Feller, Non-
[187] B. Pharse, P. Lu, T. Wipf, L. Greimann, J. Seo, Field Validation of a Statistical- intrusive health monitoring of infused composites with embedded carbon
Based Bridge Damage-Detection Algorithm, J. Bridge Eng. 18 (2013) 1227–1238. quantum piezo-resistive sensors, Compos. Sci. Technol. 123 (2016) 286–294.
[188] Q. Xia, Y.Y. Cheng, J. Zhang, F.Q. Zhu, In-Service Condition Assessment of a Long- [219] L. Gao, E.T. Thostenson, Z. Zhang, T.W. Chou, Coupled carbon nanotube network
Span Suspension Bridge Using Temperature-Induced Strain Data, J. Bridge Eng. and acoustic emission monitoring for sensing of damage development in compo-
04016124 (2016). sites, Carbon 47 (2009) 1381–1388.

89
S. Das, P. Saha Measurement 129 (2018) 68–90

[220] H.S. Park, H.G. Sohn, I.S. Kim, J.H. Park, Application of GPS to monitoring of placement in structural health monitoring, Smart Mater. Struct. 105033 (2012) p.
wind-induced responses of high-rise buildings, Struct. Des. Tall Special Build. 17 9 pp.
(1) (2008) 117–132. [225] T.H. Yi, H.N. Li, C.W. Wang, Multiaxial sensor placement optimization in struc-
[221] Z.N. Li, J. Tang, Q.S. Li, Optimal sensor locations for structural vibration mea- tural health monitoring using distributed wolf algorithm, Struct. Control Health
surements, Appl. Acoust. 65 (2004) 807–818. Monit. 23 (2016) 719–734.
[222] A.R.M. Rao, K. Lakshmi, S. Krishnakumar, A Generalized Optimal Sensor [226] T.H. Loutas, A. Bourikas, Strain sensors optimal placement for vibration-based
Placement Technique for Structural Health Monitoring and System Identification, structural health monitoring. The effect of damage on the initially optimal con-
Procedia Eng. 86 (2014) 529–538. figuration, J. Sound Vib. 410 (2017) 217–230.
[223] T.H. Yi, H.N. Li, X.D. Zhang, Sensor placement on Canton Tower for health [227] D.D. Cong, H.D. Trung, T.N. Thoi, An efficient approach for optimal sensor pla-
monitoring using asynchronous-climb monkey algorithm, Smart Mater. Struct. cement and damage identification in laminated composite structures, Adv. Eng.
125023 (2012) p. 12 pp. Softw. 119 (2018) 48–59.
[224] T.H. Yi, H.N. Li, X.D. Zhang, A modified monkey algorithm for optimal sensor

90