applied

sciences
Article
Passively Conducted Vibration Sensing with Fiber
Bragg Gratings
Lipi Mohanty 1 , Yaowen Yang 1, * and Swee Chuan Tjin 2, *
1 School of Civil and Environmental Engineering, Nanyang Technological University, 50 Nanyang Avenue,
Singapore 639798, Singapore; lipi@ntu.edu.sg
2 School of Electrical and Electronic Engineering, Nanyang Technological University, 50 Nanyang Avenue,
Singapore 639798, Singapore
* Correspondence: cywyang@ntu.edu.sg (Y.Y.); esctjin@ntu.edu.sg (S.C.T.); Tel.: +65-6790-5279 (Y.Y.);
+65-6790-4845 (S.C.T.)




Received: 14 July 2018; Accepted: 28 August 2018; Published: 10 September 2018


Featured Application: Fiber sensor to measure vibrational frequency of civil structures.

Abstract: Measuring vibrations is a common method of monitoring the integrity of structures and
heavy machinery, that are subject to dynamic loads. Strong vibrations for prolonged periods of
time can be caused by various sources, such as trains, motors and heavy machinery. These strong
vibrations should be identified and managed to ensure operational safety. This study proposes a
flexible metal beam sensor with a fiber Bragg grating (FBG) mounted on the surface to measure
the vibrational frequency. We present a sensor for measuring the vibrational frequencies on-site by
placing the beam so that it makes physical contact with the vibrating body. The sensor has been
tested in the range of 50–200 Hz. The sensing beam can detect the vibrations that are induced in
other metallic bodies where there are metallic structures of low stiffness to conduct the vibration.
The results show that the sensing beam is capable of detecting the frequency of forced vibrations
from its periphery when placed in different orientations.

Keywords: fiber Bragg gratings; vibration sensors; forced vibration

1. Introduction
It is routine to obtain vibration measurements in the structural health monitoring of suspension
bridges, railway tracks, high-rise buildings etc. Vibrations near a railway track can be a matter of
concern [1]. Especially due to the high speed trains [2] travelling at speeds in excess of 200 km/h
or even 300 km/h, nearby buildings are likely to be affected. When the train tracks are built in
cities, whether they are above ground or underground, they generate a certain amount of vibration
in nearby buildings and structures [3,4]. Such vibrations at low levels can be acceptable unless there
is resonance. Resonance causes damage to buildings and can make them unsafe or unstable. Thus,
certain international standards have recommended regular quantification and management of these
vibrations [5], especially those up to 250 Hz. It is important to monitor the vibration frequencies as
they often act a benchmark for normal operations. However, it may be difficult to attach electrical or
bulky sensors to railway tracks due to safety concerns. Assembling equipment or accessories near the
tracks for obtaining measurements is also a challenge. Hence, fiber Bragg grating (FBG) sensors have
been found to hold great promise in this area [6,7] since the fiber cables can run for several meters
and allow for continuous and remote monitoring. In particular, in train tunnels, there is very little
space for workers to conduct checks with conventional accelerometers when the lines are in operation.
However, with fiber sensors, the system can be used without disrupting normal operations.

Appl. Sci. 2018, 8, 1599; doi:10.3390/app8091599 www.mdpi.com/journal/applsci

Appl. Sci. 2018, 8, 1599 2 of 12

Fiber Bragg gratings (FBG) have been used as vibration sensors for nearly two decades.
The conventional use of FBGs for vibration measurements involves gluing the bare FBG directly
on to the vibrating object [8,9]. Since the FBG is fragile and establishing good contact is crucial, this
method is most suitable in laboratory conditions. Some alternate, packaged designs of FBG-based
vibration sensors are also commercially available [10,11]. The FBG sensor has the main advantage of
intrinsic sensitivity without additional complex packaging or design. In this paper, we present a sensor
that is configured as a simple rectangular beam with a uniform FBG and extrapolates the concept of a
cantilever. Cantilever vibration and modal studies with a FBG sensor have been reported by several
groups [12,13]. The modal frequency varies depending on the configuration and dimensions of the
cantilever. The results in literature have mainly focused on the shift in modal frequencies.
However, the objective of this study is to detect the vibration frequencies, apart from the modal
frequencies, that affect the beam at a given time. This paper presents the results from a passive
vibration sensor. The sensor design includes a rectangular metal beam of low stiffness bearing a
fiber Bragg grating, which detects vibration frequencies from the vibrating object that it is coupled
to. It is based on the principle of the coupling of forced vibrations. The FBG measures two types of
frequencies due to the beam vibration: (a) the modal frequencies of the beam, which can be calculated
numerically and (b) other frequencies from the periphery due to forced vibrations. A rectangular
beam of a certain mass and dimension has a fixed set of vibration modes and is likely to exhibit those
modal frequencies [14,15]. However, other frequencies measured by this sensor are likely to be from its
peripheral environment or nearby sources of vibration. This sensor relies on the coupling of vibrations
from one metallic object to the other. Measuring vibration frequencies is a proven method of detecting
structural integrity. This vibration sensor can be used to conduct intermittent checks on various types
of structures and machinery by passively conducting the vibrations from them.

2. Experimental Methods
A FBG with a central wavelength at 1549 nm, which was fabricated on a single-mode fiber, was
used as the sensor. The bare FBG was glued to a thin and light metal beam (aluminum, 470 mm ×
15 mm × 0.3 mm) with a rectangular cross-section. The FBG was located at around 15 cm from one end
of the beam. The wavelength shift was recorded with the Micron optics interrogation (MOI) system
sm130 (Micron Optics, Atlanta, GA, USA). The sampling frequency of the interrogation system was
limited to 500 Hz but the same measurements can be obtained using an edge filter and a photoreceiver
or using alternate interrogation systems with higher sampling frequency. Two common acoustic sources
were used for the experiments in this study. An LG phone was used in-contact for pulsed vibrations
and a Bluetooth speaker (connected to a phone) was used for frequency tones. The wavelength shift
was recorded in the time domain and converted to the frequency domain by Fast Fourier Transform
(FFT) function in MATLAB (R2014a, MathWorks, Natick, MA, USA). The sensor beam was tested
in two different configurations. The first batch of experiments was set up with another beam acting
as a wedge (or intermediate medium) that couples vibrations on to the sensing beam. The tested
configurations are described as cases 1 to 4. The second setup used a simply supported aluminum
plate as the intermediate medium, with the sensing beam placed near one edge on top of the plate
(case 5). Finally, the field application test is described in case 6, where the sensing beam was propped
against the external casing of an operational three-phase motor.
The following is the list of cases tested with the sensing beam to determine whether it is able to
detect forced vibration frequencies in different orientations.

A. First setup: Beam on beam

(i) Case 1: Test with different frequencies from a speaker

(ii) Case 2: Test with phone vibration simultaneously with beam oscillation
(iii) Case 3: Test with the beam moved closer to speaker
(iv) Case 4: Test with the FBG bearing surface flipped downwards
Appl. Sci. 2018, 8, 1599 3 of 12

B. Second setup: Beam on plate

(v) Case 5: Test with different frequencies and comparison with analytical results
Appl. Sci. 2018, 8, x FOR PEER REVIEW 3 of 13
C. Third setup: Beam propped against operational motor
C. Third setup: Beam propped against operational motor
(vi) Case 6: Beam in contact with motor housing
(vi) Case 6: Beam in contact with motor housing
Figure 11 shows
Figure shows one
one of of the
the configurations
configurations used used to
to test
test the
the sensor
sensor (case
(case 1)
1) for
for detecting
detecting different
different
vibrational
vibrational frequencies. A metallic block was placed on an optical table. A beam
A metallic block was placed on an optical table. A beam with a length with a length of of
15
cmcm
15 andandthickness of 1ofmm
thickness 1 mm waswasplaced with
placed oneone
with endendon the
on theblock andand
block the the
other endend
other free,free,
which is a
which
similar
is position
a similar to cantilevers.
position to cantilevers. Beam
Beam 1 was
1 was weighed
weigheddown downwith withan anacoustic
acoustic speaker,
speaker, which was was
termed the
termed the fixed
fixed end.
end. The
The speaker
speaker waswas placed
placed face
face down
down to to allow
allow better
better coupling
coupling ofof the
the sound
sound toto the
the
metal. The
metal. The sensing
sensing beam
beam was was placed
placed without
without fasteners
fasteners on on the
the free
free end
end of
of beam
beam 1.1. Several orientations
orientations
of this
of this configuration
configuration were
were tested
tested with
with the
the ends
ends of
of the beam 2: both taped-free; both touching the table;
or one
or one end
end touching
touching the
the table
table and
and the
the other
other end
end free
free to
to oscillate.
oscillate. The
The FBG
FBG was
was either
either attached
attached toto the
the
wedge point
wedge point oror within
within 55 cmcm distance
distance of
of it.
it. All
All the
the tests
tests were
were repeated
repeated several
several times
times and
and the
the selected
selected
cases were
cases were analyzed
analyzed inin the
the next
next section.
section.

Figure 1.1.Experimental
Figure Experimentalconfiguration
configuration with
with oneone beam
beam bearing
bearing the the
FBGFBG
(fiber(fiber
BraggBragg grating)
grating) and
and the the
other
other conducting
beam beam conducting the vibration
the vibration from the from
acousticthesource.
acoustic source.
Both Both
ends of beam ends ofsimply
2 are beam supported
2 are simply
by
supported
the by schematic
table. The the table. The schematic
is shown is left.
on the shown on the left.

Figure
Figure22shows
showsthe theschematic
schematic forfor
case
case2 where a phone
2 where is replaced
a phone as theasacoustic
is replaced source.source.
the acoustic The setup
The
is used for testing the vibrations of a phone, which is superposed on the oscillations
setup is used for testing the vibrations of a phone, which is superposed on the oscillations of beam 2. of beam 2. Figure 3
shows
Figurethe picturethe
3 shows and the schematic
picture and the used for case
schematic 3, where
used the3,sensing
for case beam
where the 2 was beam
sensing moved2 closer to the
was moved
Bluetooth speaker
closer to the (from speaker
Bluetooth B to A). In this B
(from measurement
to A). In thisconfiguration,
measurementthe sensing beamthe
configuration, wassensing
placed beam
a few
cm away from the free end. Since the sensing beam is pliant and flexible, this
was placed a few cm away from the free end. Since the sensing beam is pliant and flexible, this configuration introduces a
higher tensile pre-strain
configuration introduces inathe FBG, tensile
higher which causes the baseline
pre-strain to shift
in the FBG, to acauses
which higherthe
wavelength.
baseline to Inshift
case to
4,
similar
a highertests were conducted
wavelength. bysimilar
In case 4, flipping thewere
tests sensor aimed toby
conducted ascertain
flippingwhether theaimed
the sensor orientation of the
to ascertain
FBG-bearing side changed
whether the orientation of the response
FBG-bearing to vibrations.
side changed the response to vibrations.
Figure 4 shows the experimental configuration of case 5, which was conducted with the sensing
beam placed on a flat aluminum plate. Furthermore, this beam was simply supported along two
edges. The speaker was placed near one edge and the sensor beam was placed near the other
edge. The distance between the edge of the speaker and the sensing beam was more than 10 cm.
These laboratory tests were followed by field tests where the sensor was tested with an operational
three-phase motor (case 6).
 setup is used for testing the vibrations of a phone, which is superposed on the oscillations of beam 2.
Figure 3 shows the picture and the schematic used for case 3, where the sensing beam 2 was moved
closer to the Bluetooth speaker (from B to A). In this measurement configuration, the sensing beam
was placed a few cm away from the free end. Since the sensing beam is pliant and flexible, this
configuration introduces a higher tensile pre-strain in the FBG, which causes the baseline to shift to
Appl.aSci.
higher
2018, wavelength.
8, 1599 In case 4, similar tests were conducted by flipping the sensor aimed to ascertain
4 of 12
whether the orientation of the FBG-bearing side changed the response to vibrations.

Appl. Sci. 2018, 8, x FOR PEER REVIEW 4 of 13

Figure 2. The schematic shows phone replacing the speaker, with top end of sensing beam lightly
taped and bottom end allowed to oscillate. The sensing beam 2 was made to oscillate by tapping.

Appl. Sci. 2018, 8, x FOR PEER REVIEW 4 of 13

Figure 2. The
Figure schematic
2. The shows
schematic showsphone
phonereplacing the speaker,
replacing the speaker,with
withtop
topend
end
of of sensing
sensing beambeam lightly
lightly
taped andand
taped bottom endend
bottom allowed totooscillate.
allowed oscillate.The
The sensing beam22was
sensing beam wasmade
madetoto oscillate
oscillate by by tapping.
tapping.

Figure 3. Experimental configuration with sensing beam moved closer to the acoustic source. The
schematic shows positions A and B.

Figure 4 shows the experimental configuration of case 5, which was conducted with the sensing
beam placed on a flat aluminum plate. Furthermore, this beam was simply supported along two
edges. The speaker was placed near one edge and the sensor beam was placed near the other edge.
The distance between the edge of the speaker and the sensing beam was more than 10 cm. These
Figure
Figure 3. Experimental configuration
Experimental
3. tests configuration with
withsensing
sensingbeam moved
beam closercloser
moved to the to
acoustic source. The
theoperational
acoustic source.
laboratory were followed by field tests where the sensor was tested with an three-
schematic
The schematic shows
showspositions A and B.
phase motor (case 6). positions A and B.

Figure 4 shows the experimental configuration of case 5, which was conducted with the sensing
beam placed on a flat aluminum plate. Furthermore, this beam was simply supported along two
edges. The speaker was placed near one edge and the sensor beam was placed near the other edge.
The distance between the edge of the speaker and the sensing beam was more than 10 cm. These
laboratory tests were followed by field tests where the sensor was tested with an operational three-
phase motor (case 6).

Figure
Figure 4. Experimental
4. Experimental configuration
configuration with
with beam
beam bearing
bearing thethe FBG
FBG placed
placed onon a thin
a thin metal
metal plate.The
plate. Theplate
was simply supported along two edges. Speaker was placed at a distance of more than 10 cmcm
plate was simply supported along two edges. Speaker was placed at a distance of more than 10 away
away from the beam. The schematic shows that the sensing beam 2 is placed on the metal plate
from the beam. The schematic shows that the sensing beam 2 is placed on the metal plate without any
without any fasteners and the speaker is also placed on the plate.
fasteners and the speaker is also placed on the plate.
3. Results
3. Results andand Discussion
Discussion
The results obtained from the different cases bearing
are presented
the FBGinplaced
this section.
a thinFor case 1 (Figure 1),
TheFigure 4. obtained
results Experimental configuration
from with cases
the different beam are presented in thisonsection. metal
For plate.
case 1The
(Figure 1),
the speaker was
plate was used
was simply to play sine tones of different frequencies to generate acoustic vibrations. The
the speaker used supported alongtones
to play sine two edges. Speaker was
of different placed at a to
frequencies distance of more
generate than 10 cm
acoustic vibrations.
time-dependent
away from theresponse of the
beam. The wavelength
schematic showsshift
that of
thethe FBG beam
sensing was recorded
2 is placedwith themetal
on the MOI.plate
Figure 5
The shows
time-dependent
the FFT
response of the wavelength shift of the FBG was recorded with the MOI. Figure 5
without anyoffasteners
the wavelength shift data.
and the speaker is alsoInplaced
this case, the
on the frequencies of 50 Hz, 80 Hz and 110 Hz
plate.
showswere swept one after the other after using the speaker for different periods of time. The difference110
the FFT of the wavelength shift data. In this case, the frequencies of 50 Hz, 80 Hz and in Hz
werethe
3.swept
number
Results oneofafter
and the other
samples
Discussionresultsafter using the
in different speaker for
magnitudes different
although periods
all three of time.are
frequencies The difference
detected.
in the number of samples results in different magnitudes although all three
The sensing beam oscillation is negligible in comparison throughout the three-frequency sweep, frequencies are detected.
The results obtained from the different cases are presented in this section. For case 1 (Figure 1),
which is shown as the small peak under 10 Hz. Due to the experimental arrangement and flexibility
the speaker was used to play sine tones of different frequencies to generate acoustic vibrations. The
of the beam 2, the ends are touching the table. However, the FBG detects the forced vibrations.
time-dependent response of the wavelength shift of the FBG was recorded with the MOI. Figure 5
shows the FFT of the wavelength shift data. In this case, the frequencies of 50 Hz, 80 Hz and 110 Hz
were swept one after the other after using the speaker for different periods of time. The difference in
the number of samples results in different magnitudes although all three frequencies are detected.
Appl. Sci. 2018, 8, 1599 5 of 12

The sensing beam oscillation is negligible in comparison throughout the three-frequency sweep, which
is shown as the small peak under 10 Hz. Due to the experimental arrangement and flexibility of the
Appl. Sci. 2018, 8, x FOR PEER REVIEW 5 of 13
beam 2, Sci.
Appl. the2018,
ends 8, xare
FORtouching the table. However, the FBG detects the forced vibrations. 5 of 13
PEER REVIEW

6
6
3 frequency sweep
3 frequency sweep
5
5

4
(a.u.)

4
Magnitude(a.u.)

3
Magnitude

2
2

1
1

0
00 50 100 150
0 50 Frequency (Hz) 100 150
Frequency (Hz)
Figure 5. FFT
5. FFT (Fast
(Fast FourierTransform)
Fourier Transform) data of
of consecutively swept frequencies at 50
atHz, 80 Hz80and
Figure
Figure 5. FFT (Fast Fourier Transform)data
data of consecutively swept
consecutively swept frequencies
frequencies at 50 50
Hz,Hz,
80 Hz Hz
and and
110
110 Hz Hz
withwith random
random number
number of
of samples.
samples. Small
Small peak
peak observed
observed near
near7 Hz
7 Hzdue to
due natural
to oscillations.
natural oscillations.
110 Hz with random number of samples. Small peak observed near 7 Hz due to natural oscillations.

Figure
Figure 6 shows
6 shows thethe time-dependent
time-dependent oscillationofofthe
oscillation thesensing
sensingbeam beam withthe the pulsed vibrations
vibrations
Figure 6 shows the time-dependent oscillation of the sensing beamwith with thepulsed
pulsed vibrations that
that are superposed. The oscillation was generated by gentle tapping of the free end of beam 2 (Figure
are superposed. The oscillation
that are superposed. was generated
The oscillation was generated by gentle tapping
by gentle tapping ofof the
thefree
freeend
end of beam22(Figure
of beam (Figure 2).
2). In this case (case 2), the wedge end had a phone vibrating in beeps. The vibration was
In this
2). case (case
In this 2), (case
case the wedge
2), theend had aend
wedge phone hadvibrating
a phone in beeps. The
vibrating vibration
in beeps. The was automatically
vibration was
automatically superposed on the cantilever oscillation although there was a gap of several cm on
automatically
superposed on the superposed
cantilever on the cantilever
oscillation although oscillation
there was although
a gap ofthere
several wascm a on
gapbeam
of several
1. The cm on
vibration
beam 1. The vibration was applied in pulses due to the pattern of beeps selected as the cantilever was
beam 1.
wasoscillating. The
applied inWhen vibration
pulses due was applied
to the patternin pulses
of at due
beeps to the pattern
selected of beeps
as the cantilever selected as the cantilever
wasisoscillating. was
When
the cantilever oscillates its natural frequency, the vibration at frequencies of the
oscillating.
cantilever Whenat
oscillates the
itscantilever
natural oscillates atthe
frequency, its vibration
natural frequency,
is at the vibration
frequencies of is at frequencies
around 7–15 Hz, of
which
around 7–15 Hz, which depends on the length of the beam that is unconstrained. However, the
around
depends 7–15 Hz,
on thephone
length which depends
of the beam on the
that150 length of
is unconstrained. the beam that
However, is unconstrained.
the superposed However,
phone the
vibration
superposed vibration is around Hz. The forced vibration is detected by the sensor although
superposed phone vibration is around 150 Hz. The forced vibration is detected by the sensor although
is around
it occurs 150 Hz. The
together forced
with the lowvibration
frequencyis detected
oscillation. byThe
the phone
sensorvibration
althoughwas it occurs together
identified with the
separately
it occurs together with the low frequency oscillation. The phone vibration was identified separately
lowwith a National
frequency Instruments
oscillation. (NI, Austin,
The phone vibrationTX, wasUSA)identified
accelerometer connected
separately with to NI SignalExpress
a National Instruments
with a National Instruments (NI, Austin, TX, USA) accelerometer connected to NI SignalExpress
(NI,system
Austin,(2012). The PSD
TX, USA) (power spectrum
accelerometer connected distribution) showed thatsystem
to NI SignalExpress
system (2012). The PSD (power spectrum distribution) showed that the peak frequency of the
the peak frequency
(2012). The PSD of the
(power
vibration
spectrum beeps was
distribution) near
showed 150 Hz.
that the peak frequency of the vibration beeps was near 150 Hz.
vibration beeps was near 150 Hz.

1549.68
1549.68

1549.66
1549.66

1549.64
1549.64
Wavelength (nm)
(nm)

1549.62
Wavelength

1549.62

1549.6
1549.6

1549.58
1549.58

1549.56
1549.56
6000 6500 7000 7500 8000 8500 9000 9500 10000
6000 6500 7000 7500 8000 (msec)8500
Time samples 9000 9500 10000
Time samples (msec)
Figure 6. Time varying data of central wavelength of FBG, showing that the damped oscillation of
Figure 6. Time varying data of central wavelength of FBG, showing that the damped oscillation of the
Figure 6.beam
the sensing
sensing Time varying
behavesdata ofcantilever.
central wavelength of FBG, showing from
that the damped oscillation of the as
beam behaves asasa acantilever. The
The pulsed
pulsed vibrations
vibrations from the the phone
phone are are superposed
superposed as
sensing
periodic highbeam behaveswaves.
frequency as a cantilever.
The The
x-axis pulsed
scale is in vibrationsoffrom
samples time the phone are superposed as
(ms).
periodic high frequency waves. The x-axis scale is in samples of time (ms).
periodic high frequency waves. The x-axis scale is in samples of time (ms).
Appl. Sci. 2018, 8, x FOR PEER REVIEW 6 of 13
Appl.
Appl. Sci.
Sci. 2018,
2018, 8,
8, x1599
FOR PEER REVIEW 66 of 13
of 12
Figure 7 shows the FFT of the time-domain signal that is plotted in Figure 6. In this case, the
relative magnitude
Figure 7 showsofthe theFFTlowof
frequency peak duesignal
the time-domain to oscillation is higherinthan
that is plotted the 6.
Figure forced
In thisvibration
case, the at
150 Hz
relative (phone
magnitude vibration).
of the lowHowever,
frequency when
peak the
due cantilever
to oscillation
oscillation is higher dies
than
Figure 7 shows the FFT of the time-domain signal that is plotted in Figure 6. In this case, the down
the and
forced the pulsed
vibration at
vibration
150
relative of the phone
Hz magnitude
(phone continues
vibration).
of the low (Figure
However,
frequency 8), the
when
peak theFFT
due todata
cantilever shows that
oscillation
oscillation is the dies
higher 150
thanHz peak
down
the from
and
forced thethepulsed
vibrationphone at
is the
vibrationdominant
of the peak
phone (Figure
continues 9). The
(Figure section
8), theof the
FFT time
data spectrum
shows that that
the is
150
150 Hz (phone vibration). However, when the cantilever oscillation dies down and the pulsed vibration selected
Hz peak in Figure
from the 8 is
phone the
part
is
of thewhere
the phonethe
dominant phone vibration
peak (Figure
continues (Figure isThe
9).8), dominant.
thesection
FFT data Due
of the totime
shows thethat
lowthe
stiffness
spectrum 150that ofispeak
Hz the sensing
selected
frominthe beam,
Figure
phone the ishigh
8 is the
the
frequency
part
dominant vibrations
wherepeak phoneare
the (Figure 9). measured
vibration
The section although
is dominant.
of the timethere
Due toisthe
spectrum no obvious
low or visible
thatstiffness
is selected of the oscillation
sensing
in Figure of the
8 isbeam,
the sensing
the
part high
where
beam.
frequency
the phone vibrations
vibration is are measuredDue
dominant. although
to thethere is no obvious
low stiffness of theor visiblebeam,
sensing oscillation
the highof the sensing
frequency
beam.
vibrations are measured although there is no obvious or visible oscillation of the sensing beam.
10
109

98

87
(a.u.)

76
(a.u.)
Magnitude

65
Magnitude

54

43

32

21

10
0 50 100 150 200 250
0 Frequency (Hz)
0 50 100 150 200 250
Frequency (Hz)
Figure 7. FFT data of cantilever vibration with a smaller peak at 150 Hz due to the phone.
Figure FFT data
7. FFT
Figure 7. data of
of cantilever
cantilever vibration
vibration with
with aa smaller
smaller peak
peak at
at 150
150 Hz
Hz due
due to
to the
the phone.
phone.

1549.68
1549.68
1549.66
1549.66
1549.64
(nm)

1549.64
(nm)

1549.62
Wavelength

1549.62
Wavelength

1549.6
1549.6
1549.58
1549.58
1549.56
1549.56
1.3 1.4
1.6 1.5
1.7 1.8 1.9 2
Time samples (msec) x 102
4
1.3 1.4 1.5 1.6 1.7 1.8 1.9
Time samples (msec) x 10
4
Figure
Figure 8. Time varying
8. Time data of
varying data of central
central wavelength position of
wavelength position of FBG
FBG where
where the
the phone
phone vibration
vibration is is
dominant.
dominant. There was
was no
Therevarying oscillation
nodata
oscillation and the
the sensing
sensing beam
and wavelength beam appeared
appeared to be
be static. The configuration is
Figure 8. Time of central position of FBGtowhere
static.
theThe configuration
phone vibration is is
the
the same
same as shown
asThere
shown in Figure
in no
Figure 22 schematic.
schematic.
dominant. was oscillation and the sensing beam appeared to be static. The configuration is
the same as shown in Figure 2 schematic.
Appl. Sci. 2018, 8, 1599 7 of 12
Appl. Sci. 2018, 8, x FOR PEER REVIEW 7 of 13
Appl. Sci. 2018, 8, x FOR PEER REVIEW 7 of 13
10
10
9
9
8
8
7
7
Magnitude(a.u.)

6
Magnitude(a.u.)

6
5
5
4
4
3
3
2
2
1
1
0
00 50 100 150 200 250
0 50 100 Frequency (Hz) 150 200 250
Frequency (Hz)
Figure9.9.FFT
Figure FFTdata
dataof
ofnegligible
negligible cantilever vibrations with
cantilever vibrations with aa dominant
dominantpeak
peakatat150
150Hz
Hzdue
duetotothe
thephone.
phone.
Figure 9. FFT data of negligible cantilever vibrations with a dominant peak at 150 Hz due to the phone.
Figure
Figure 3 shows
3 shows thethe
measurement
measurement configuration
configuration used,
used,with thethe
with sensing
sensing beambeam placed
placed a few cm cm
a few away
from Figure
the free 3 shows
end (case the
3). measurement
This configurationconfiguration
introduces used,
a with the
higher sensing
tensile beam placed
pre-strain in the a few which
FBG, cm
away from the free end (case 3). This configuration introduces a higher tensile pre-strain in the FBG,
awaythe
causes from the freetoend
baseline (case
shift 3). This
(Figure 10)configuration
to a10) higher introduces a higher
wavelength tensile pre-strain
by approximately 0.020.02 in
nm.theTheFBG, two
which causes the baseline to shift (Figure to a higher wavelength by approximately nm. The
which
spectra causes
markedmarkedthe baseline
A and A to
B denote shift (Figure
the wavelength10) to a higher wavelength
shift measured by approximately
at positions A andABand 0.02
of thenm. The
sensing
two spectra and B denote the wavelength shift measured at positions B of the
two respectively.
beam, spectra marked A and
When B80denote
anWhen the wavelength
Hzansignal shift
testmeasured atwith
positions A and Bthe of lower
the
sensing beam, respectively. 80 Hzwas signalusedwastoused both
to testcases
both cases the speaker,
with the speaker, the
sensinghad
baseline beam, a respectively.
relatively lowerWhenpeak anthan
80 Hz thesignal
higherwasbaseline
used to (Figure
test both11).cases with
This maythebe speaker,
because thethe
lower baseline had a relatively lower peak than the higher baseline (Figure 11). This may be because
lower baseline had a relatively lower peak than the higher baseline (Figure 11). This may be because
the sensor
sensor was effectively
was effectively moved moved
closercloser
to thetoacoustic
the acoustic
sourcesource and received
and received a higher
a higher amplitude.
amplitude. TheThe same
the sensor was effectively moved closer to the acoustic source and received a higher amplitude. The
same number
number of samples of samples
was used wastoused to calculate
calculate the FFT theinFFT
bothin cases.
both cases. Normalizing
Normalizing the the
twotwo signals
signals with
same number of samples was used to calculate the FFT in both cases. Normalizing the two signals
with reference to the central wavelength also yielded similar spectra. In
reference to the central wavelength also yielded similar spectra. In this case, there is also no oscillationthis case, there is also no
with reference to the central wavelength also yielded similar spectra. In this case, there is also no
of oscillation
the beam and of thenobeam
peakand no peak
located located
around around
10 Hz. These10 Hz.
testsThese
havetests
beenhave been successfully
successfully repeatedrepeated
in several
oscillation of the beam and no peak located around 10 Hz. These tests have been successfully repeated
in several configurations
configurations and withtypes
and with different different types ofand
of wedges wedges and
plates. Whenplates.theWhen
gap is theingap
theisrange
in theof range
10 cm,
in several configurations and with different types of wedges and plates. When the gap is in the range
of 10 cm,
theofsensor the sensor can detect frequencies up to 200 Hz. These results prove that the sensor can be
10 cm,can thedetect
sensorfrequencies up to 200 Hz.
can detect frequencies up toThese results
200 Hz. prove
These thatprove
results the sensor
that the cansensor
be attached
can be to
attached
vibrating to vibrating structures with metal conduits to measure the vibrational frequencies. All the
attachedstructures
to vibrating with metal conduits
structures with metal to measure
conduits to themeasure
vibrational frequencies.
the vibrational All the beams
frequencies. All the and
beams and plates used were made of aluminum. Beam 1 was also tested as a sensing beam but its
plates used were made of aluminum. Beam 1 was also tested as a
beams and plates used were made of aluminum. Beam 1 was also tested as a sensing beam but its sensing beam but its stiffness was
stiffness was higher and it was not sensitive to the forced vibrations. Thus, we may conclude that
higher andwas
stiffness it was not sensitive
higher and it wastonot thesensitive
forced vibrations.
to the forced Thus, we mayThus,
vibrations. conclude
we may thatconclude
increasing thatthe
increasing the thickness or reducing the length can contribute to higher stiffness. Furthermore, when
thickness or reducing the length can contribute to higher stiffness. Furthermore,
increasing the thickness or reducing the length can contribute to higher stiffness. Furthermore, when when beam 2 was
beam 2 was completely flat on the optical table and the speaker was placed on top of it at 10 cm from
completely
beam 2 was flat on the optical
completely flat ontable and the
the optical speaker
table and the was placed
speaker was onplaced
top ofonit top
at 10of cm
it atfrom
10 cmthe from FBG
the FBG location, the sensitivity was zero. Thus, it is important to place the sensing beam in a position
location,
the FBGthe sensitivity
location, was zero. was
the sensitivity Thus, it isThus,
zero. important to place to
it is important the sensing
place beam inbeam
the sensing a position where it
in a position
where it can flex in both directions to allow for the presence of compressive and tensile stresses.
canwhere
flex init both
can flex in both directions
directions to allow for to allow for the presence
the presence of compressive
of compressive and tensile andstresses.
tensile stresses.
1549.57
1549.57

1549.56
1549.56
A
(nm)

1549.55 A
(nm)

1549.55
Wavelength
Wavelength

1549.54
1549.54 B
B
1549.53
1549.53

1549.52
1549.52 2000 2500 3000 3500 4000
2000 2500 Time samples (msec)
3000 3500 4000
Time samples (msec)

Figure
Figure 10.10. Timevarying
Time varyingdata
dataof
ofcentral
central wavelength
wavelength position
positionofofFBG
FBG(beam
(beam2)2)atat
two different
two locations
different locations
Figure 10. Time varying data of central wavelength position of FBG (beam 2) at two different locations
along beam 1 due to the 80 Hz tone. The lower signal in the data is measured at position
along beam 1 due to the 80 Hz tone. The lower signal in the data is measured at position B and B and upper
upper
along beam 1 due to the 80 Hz tone. The lower signal in the data is measured at position B and upper
signal
signal at at position
position A,A,
asas shown
shown inin the
the schematic
schematic of
of Figure
Figure 3.
3.
signal at position A, as shown in the schematic of Figure 3.
Appl. Sci. 2018, 8, 1599 8 of 12
Appl. Sci. 2018, 8, x FOR PEER REVIEW 8 of 13
Appl. Sci. 2018, 8, x FOR PEER REVIEW 8 of 13
3
3 A
A
B
2.5 B
2.5

(a.u.)
2

Magnitude
Magnitude (a.u.) 1.5
1.5

1
1

0.5
0.5

0
00 50 100 150
0 50 Frequency (Hz) 100 150
Frequency (Hz)

Figure11.11. FFTdata
data of 80Hz
Hz tone measured
measured atat two
two different
differentdistances
distancesfrom
fromthe
theacoustic source,
Figure
Figure 11.FFT
FFT dataofof80
80 Hz tone
tone measured at two different distances from the acoustic
acoustic source,
source,
positions
positions A A and
and B,B,
asas shown
shown in
in schematic
schematic of
of Figure
Figure 3.
3.
positions A and B, as shown in schematic of Figure 3.
Figure
Figure 1212 showsthe
shows the FFTdata
data ofthree
three tests (case
(case 4).
4). A
A frequency
frequencytone
toneof 150
150Hz was used to to
test
Figure 12 shows theFFT
FFT dataofof threetests
tests (case 4). A frequency tone ofof150 HzHz was
was used
used to testtest
the
thethe effect
effect of
of of vibrations
vibrations when the FBG bearing surface was facing upwards with a central wavelength
effect vibrationswhen
whenthe
theFBG
FBGbearing
bearing surface was facing
surface was facingupwards
upwardswith witha acentral
central wavelength
wavelength
that is near 1549.72 nm. After this, the sensing beam was flipped at the same location and the central
that
that is near 1549.72 nm. After this, the sensing beam was flipped at the same location and thethe
is near 1549.72 nm. After this, the sensing beam was flipped at the same location and central
central
wavelength of the grating was shifted to 1549.68 nm. However, the repeated measurements showed
wavelength
wavelength ofof
the grating
the gratingwas
wasshifted
shiftedto to1549.68
1549.68 nm. However,the
nm. However, therepeated
repeatedmeasurements
measurements showed
showed
that the forced vibrations were detected with a similar magnitude.
that thethe
that forced
forcedvibrations
vibrationswere
weredetected
detectedwith
withaa similar
similar magnitude.
magnitude.

Figure FFTFFT
12.12.
Figure data of 150
data Hz tone
of 150 usedused
Hz tone to test
tosensing beam beam
test sensing at oneatlocation, with FBG
one location, withfacing
FBG upwards.
facing
Figure
This was 12. FFT data
followed by of 150 Hz
repeated tone
tests used
with FBG to facing
test sensing beam atThe
downwards. oneforced
location, with FBG
vibration facing
was detected
upwards. This was followed by repeated tests with FBG facing downwards. The forced vibration was
upwards. This was followed by repeated tests with FBG facing downwards. The forced vibration was
although
detectedthe sensor the
although wassensor
flipped in flipped
was the z-direction.
in the z-direction.
detected although the sensor was flipped in the z-direction.
WeWeoutline
outline the possible
the possiblereasons
reasonsforforthe
thesensor
sensor response inin the
thefollowing
followingparagraph.
paragraph.AsAs thethe length
length
We outline the possible reasons for the sensor response in the following paragraph. As the length
of of
thethebeam
beamisissignificantly greaterthan
significantly greater thanthe the width
width andand thickness,
thickness, the stiffness
the stiffness of theofbeamthe isbeam
low. isThe low.
of the beam is significantly greater than the width and thickness, the stiffness of the beam is low. The
Thesensing
sensing beam
beam does
doesnot
notbehave
behave asas
a single
a single uniformly
uniformly strained
strainedcantilever
cantileverbutbut
as as
several
several sections
sections thatthat
sensing beam does not behave as a single uniformly strained cantilever but as several sections that
behave
behave asas smallercantilever-type
smaller cantilever-typesections.
sections. When the the sensor
sensorbeam
beamwaswaswedged
wedgednear near thethecenter
center and and
behave as smaller cantilever-type sections. When the sensor beam was wedged near the center and
waswas free
free to to vibrate
vibrate ononboth
bothends,
ends,dual
dualpeaks
peakswere
were obtained
obtained that
that were
were both
bothunder
under2020Hz. Hz.
was free to vibrate on both ends, dual peaks were obtained that were both under 20 Hz.
The
The sensing
sensing capability
capability is possible
is possible in in
thisthis type
type of wedged
of wedged design
design because
because it creates
it creates a pre-strain.
a pre-strain. This
The sensing capability is possible in this type of wedged design because it creates a pre-strain.
This
allows allows vibrations to induce compression and extension on the fiber grating periods. Thus, its
This vibrations to induce
allows vibrations to compression and extension
induce compression on the fiber
and extension grating
on the fiber periods. Thus, itsThus,
grating periods. sensitivity
its
sensitivity to
to vibrations vibrations
increases, increases,
while the FBG while
is the FBG
unable to is unable or
compress to expand
compress or expand
freely when freely
it is when
stuck or it is
sensitivity to vibrations increases, while the FBG is unable to compress or expand freely when itglued
is
to metallic surfaces. The flexibility of the beam allows the signals to pick up multiple vibrational
Appl. Sci. 2018, 8, 1599 9 of 12

frequencies in the same region as the FBG. The acoustic sources (phone and speaker) used in the
experiments act as strong sources of forced vibrations. Thus, the theory and analysis of the sensor may
depend on the forced vibrations coupled to the flexible beam and its stiffness.
To support the above-mentioned concept, the response of the sensor can be analyzed using the
principle of forced vibrations [16]. The acoustic force is assumed to be a time-varying tone of frequency
ω f at a given amplitude F0 , which can be described as:

F = F0 sin ω f t (1)

The acoustic wave propagates through the base plate or beam 1, before the vibration is coupled to
the sensing beam.
The steady state solution [16] for such a forced vibration is given by x in the equation below. In this
case, an undamped solution is considered as there was a continuous source of the forced vibrations
(speaker) F.
1
x = A( ) sin (ω f t) (2)
1 − r2
F0
A= (3)
k
ωf
r= (4)
ω
F0 1
x= ( ) sin (ω f t) (5)
k 1 − r2
where the moment of inertia can be described as:

bh3
I= (6)
12
and the stiffness is:
3EI
k= (7)
L3
The following parameters were used for the simulation (E = 70 × 109 N/m2 , L = 0.47 m, b = 0.015 m
and h = 0.0003 m). The value of f in ω was taken as 10 Hz for all cases. The steady state waves for
150 Hz and 200 Hz were calculated for two different values of stiffness. As shown after substituting
Equations (6) and (7) in Equation (5), h and L have a significant role to play in determining the
amplitude of the steady state solution. Hence, a high stiffness results in low amplitude and this low
amplitude does not allow for sufficient flexion of the FBG to sense the vibrations at high frequencies.
Thus, a beam of low stiffness is more sensitive.
These calculated data were compared with experimental results from case 5 (Figure 4).
The experimental and simulated data were compared using the same number of samples (4000)
in Figure 13. The experimental data shows the main peak at the speaker frequency. Smaller peaks are
present at multiples of the main frequency, which shows the presence of other modes. For 150 Hz,
the amplitude was sufficiently strong to generate oscillations (near 10 Hz) in the freely suspended end
of the beam. The simulated data were obtained from the FFT of the steady state waves. The value of
F0 has been normalized to obtain the closest fit (0.2 N). The simulation was conducted separately for
150 Hz and 200 Hz.
Appl. Sci. 2018, 8, 1599 10 of 12
Appl.
Appl. Sci.Sci. 2018,
2018, 8, x8,FOR
x FOR PEER
PEER REVIEW
REVIEW 10 10
of of
13 13

2525
150Hz
150Hz expt
expt
200Hz
200Hz expt
expt
150Hz
150Hz sim
sim
2020 200Hz sim
200Hz sim
Magnitude (a.u.)
Magnitude (a.u.)

1515

1010

5 5

0 0
0 0 5050 100
100 150
150 200
200 250
250
Frequency
Frequency (Hz)
(Hz)

Figure
Figure
Figure 13.
13.13.FFT
FFTFFTdata
data ofofof
data experimental
experimental
experimental measurements
measurements
measurements atatat150
150150Hz
Hz Hzand
and
and 200
200200 Hz,
Hz,Hz,with
with
with sensing
sensing
sensing beam
beam
beam mounted
mounted
mounted
ononplate
plate as shown
as shown ininFigure
Figure 4. The
4. simulated
The simulated data
dataareare FFT results
FFT resultsof the
of thesinusoidal
sinusoidal
on plate as shown in Figure 4. The simulated data are FFT results of the sinusoidal waves at 150 waves at
waves 150
at 150 Hz
Hz Hz
and
and
and 200
200200 Hz
HzHz obtained
obtained
obtained from
from
from the analytical
thethe solution.
analytical
analytical solution. F is normalized
F00Fis0 is
solution. normalized
normalized to
to to0.2
0.20.2N
NN in
in inboth
both
both simulations
simulations
simulations for
forforthe
thethe
best
best
best fitfit
fit ofof
of the
the amplitude.
the amplitude.
amplitude.

ToTo
To validate
validate
validate the
the
the assumption
assumption
assumption that
that
that the
the
the sensor
sensor
sensor can
cancanbebe
be used
used
used toto
to detect
detect
detect vibrations
vibrations
vibrations caused
caused
caused byby
by dynamic
dynamic
dynamic
loads, the
loads, the sensor
sensor has
hasalso
also been
beentested
tested (case
(case 6)
6)onona three-phase
a three-phase motor
motor (55
loads, the sensor has also been tested (case 6) on a three-phase motor (55 kW) of 1475 rpm. Without(55kW)kW) of 1475
of 1475 rpm.
rpm. Without
Without
any
anyanyprior
prior
prior preparation
preparation
preparation of
ofof surfaces,
surfaces,
surfaces, the
thethe sensing
sensing
sensing beam
beam
beam was
was
was propped
propped
propped on
ononthe
the the motor
motor
motor housing
housing
housing (Figure
(Figure
(Figure 14)
14)
14)
to
totomeasure
measure
measure the
thethevibrations
vibrations
vibrations generated
generated
generated by
byby the
the theoperation
operation
operation of
ofof the
thethemotor.
motor.
motor. Figure
Figure
Figure 15
1515 shows
shows
shows the
thethe frequency
frequency
frequency
spectrum
spectrum
spectrum ofof
of this
this sensor.
this sensor.
sensor. AA
A peak
peak
peak was
was was identified
identified
identified atat
at 24.8
24.8 Hz,
24.8
Hz,Hz, which
which
which isis
is due
duedue toto
to the
thetherotation
rotation
rotation ofof
of the
thethe motor
motor
motor atat
at
1475 m −1 . Another peak was noted at 173.4 Hz due to the 7th harmonic, which is often observed in
1475
1475 mm −1 . Another
−1. Another peak
peak was was noted
noted atat 173.4
173.4 HzHz due
due toto the
the 7th
7th harmonic,
harmonic, which
which is is often
often observed
observed inin
motor
motorvibration
vibration [17]. Multiple
[17]. Multiple tests
testsshow
show that
thatthe
thepeak
peak atat173.4
173.4Hz Hz
motor vibration [17]. Multiple tests show that the peak at 173.4 Hz was present in all data.was
was present
present in all
in data.
all data.

Figure
Figure
Figure 14.14.
14. Sensor
Sensor
Sensor beam
beam
beam 2 used
22used
usedtoto to detect
detect
detect vibrational
vibrational
vibrational frequencies
frequencies
frequencies of a of of a which
motor,
a motor,
motor, which
which
was waswas
placed inplaced
placed in in
contact
contact
contact with
with thewith
motor the motor
thehousing. housing.
motor housing. The contact
The contact
The contact area was area
area was
near was near
thenear the location
the location
location of the FBG.
of the FBG.
of the FBG.
Appl. Sci. 2018, 8, 1599 11 of 12
Appl. Sci. 2018, 8, x FOR PEER REVIEW 11 of 13

20

15

Magnitude (a.u.)
10

0
0 50 100 150 200 250
Frequency (Hz)

15. The
FigureFigure 15.frequency spectrum
The frequency showing
spectrum thethe
showing primary peak
primary at at
peak 24.8 HzHzand
24.8 anda peak
a peakatatthe
the7th
7thharmonic
of 173.4harmonic
Hz. of 173.4 Hz.

These These
resultsresults
showshow thatFBG
that the the FBG
is an iseffective
an effective vibration
vibration sensor
sensor andand is easy
is easy to implement.However,
to implement.
However, the key property determining the FBG capability to detect vibrations is its ability to expand
the key property determining the FBG capability to detect vibrations is its ability to expand and contract.
and contract. A thin, long and flexible aluminum beam works as the appropriate conduit. Tests were
A thin,repeated
long and forflexible
50, 60, 80,aluminum
100, 110, 150,beam
180 and works as Continuous
200 Hz. the appropriatesweepsconduit.
of frequency Tests
have were repeated for
also been
50, 60, recorded.
80, 100, 110,
The results of the case study support the selection of a low-stiffness beam as the sensingrecorded.
150, 180 and 200 Hz. Continuous sweeps of frequency have also been
The results of the
material. case study
The sensor does notsupport the selection
have a complex design.of a low-stiffness
However, beam as the
it can be successfully sensing
retrofitted onmaterial.
metaldoes
The sensor beams,notcylinders and platesdesign.
have a complex to pick up peripheral
However, vibration
it can frequencies.retrofitted
be successfully The metallicon wedge
metal beams,
can and
cylinders be chosen
platestotobepick a cylindrical metallicvibration
up peripheral pipe, a solid rectangularThe
frequencies. wedge or anywedge
metallic other shape
can be thatchosen to
supports a pre-strain in the FBG by allowing deformation. This leaves room for the sensor to expand
be a cylindrical metallic pipe, a solid rectangular wedge or any other shape that supports a pre-strain in
and compress according to the vibrational frequency. Several tests have been repeated with a
the FBG by allowing
triangular wedgedeformation.
instead of beam This
1. For leaves room
the motor fornothe
test, sensor
wedge to expand and
or intermediate beamcompress
was present according
to the vibrational
as the sensing frequency.
beam was simplySeveral testsinhave
placed been
contact repeated
with withInstead
the housing. a triangular wedge
of a complex instead
design that of beam
1. For the motortotest,
is difficult no wedge
maintain, or intermediate
this sensor consists of a simple beam was present
low-stiffness beam assothe
thatsensing beam was
the configuration is simply
placed similar to that
in contact of athe
with bridge and canInstead
housing. consistentlyof a be strained.design that is difficult to maintain, this sensor
complex
consists of Modal
a simple studies related to the vibrations of the metal plate and cantilever have been carried out
low-stiffness beam so that the configuration is similar to that of a bridge and can
by several groups. However, in the case of this sensor, the boundary conditions are not limited to a
consistently be strained.
specific orientation or alignment. Thus, the (theoretical) modal analysis is not specific to the sensor
Modal
but to studies related
the application to the
or the casevibrations
study whereof the metal
a definite plate and
mounting method cantilever
may be used have been carried
to detect the out
by several groups.
vibrations. However,
In the in the case
analysis, although ω and ofLthis
maysensor, the boundarythe
have interdependence, conditions
sensor will are
havenot limited to a
similar
specificresponsivity
orientation forordifferent valuesThus,
alignment. of ωf. The theunderlying
(theoretical) principle
modal used in this type
analysis is notof specific
test is thattothe
the sensor
vibrations from the metal beam attached with FBG act as a source of
but to the application or the case study where a definite mounting method may be used to detect the forced vibrations, which are
detected by the FBG sensor. The frequency of the detected vibrations is the same as the source of the
vibrations. In the analysis, although ω and L may have interdependence, the sensor will have similar
forcing vibration. The vibrating body forces/couples its own vibration on/to the passive beam bearing
responsivity
the FBG.forThedifferent
results arevalues of ω
consistent f . The
with underlying
the general principleprinciple
of harmonic used in this
forced type [18]
vibration of test
where is that the
vibrations from the
the detected metal beam
frequency attached
is unchanged butwith FBG act as
the measured a source
amplitude of forced
varies with thevibrations,
stiffness and which are
detected by the
design. ThisFBG
FBGsensor. Thebe
sensor can frequency
applied toof the detected
many vibrations
areas requiring vibration is the same as the
measurements, source
such as of the
forcingmeasuring
vibration.the Thevibrations
vibrating nearbody
railway tracks, near drilling
forces/couples its ownsitesvibration
and on suspension
on/to thebridges.
passive beam bearing
the FBG. The results are consistent with the general principle of harmonic forced vibration [18] where
4. Conclusions
the detected frequency is unchanged but the measured amplitude varies with the stiffness and design.
This paper presents a fiber Bragg grating, which is attached to a flexible metal beam, that acts as
This FBG sensor can be applied to many areas requiring vibration measurements, such as measuring
a vibration sensor. The main concept in this study is the coupling of forced vibrations from a vibrating
the vibrations near railway tracks, near drilling sites and on suspension bridges.
metallic body to the sensing beam. Tests were conducted with the sensor placed at a certain distance
from the acoustic source, with intermediate metallic structures present to conduct the vibration. It
4. Conclusions
was also tested on a motor during operation. The test results show that the sensor can quantify
different
This papervibration
presents frequencies between
a fiber Bragg 50 Hz which
grating, to 200 Hz in multipletoorientations.
is attached This technique
a flexible metal can acts as a
beam, that
vibration sensor. The main concept in this study is the coupling of forced vibrations from a vibrating
metallic body to the sensing beam. Tests were conducted with the sensor placed at a certain distance
from the acoustic source, with intermediate metallic structures present to conduct the vibration. It was
also tested on a motor during operation. The test results show that the sensor can quantify different
vibration frequencies between 50 Hz to 200 Hz in multiple orientations. This technique can be an
optimal solution for retrofitting FBG-based vibration sensors on civil structures or for instant site tests.
Appl. Sci. 2018, 8, 1599 12 of 12

Author Contributions: Conceptualization, L.M.; Methodology, L.M. with inputs from Y.Y., S.C.T.; Validation, L.M.;
Formal Analysis, L.M. with feedback from Y.Y.; Investigation, L.M.; Resources, Y.Y. and S.C.T.; Data Curation, L.M.;
Writing, L.M.; Writing-Review & Editing, L.M. with significant inputs from S.C.T.; Supervision, Y.Y. and S.C.T.
Funding: This research received no external funding.
Acknowledgments: The authors would like to acknowledge the help of Deepesh Upadrashta in measuring the
frequency of phone vibration with an accelerometer.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. ISO 14837-1:2005. Mechanical Vibration—Ground-Borne Noise and Vibration Arising from Rail Systems—Part
1: General Guidance. Available online: https://www.iso.org/standard/31447.html (accessed on 17 May 2018).
2. Kogut, J.P.; Lombaert, G.; Francois, S.; Degrande, G.; Haegeman, W.; Karl, L. High Speed Train Induced
Vibrations: In Situ Measurements and Numerical Modelling. In Proceedings of the Tenth International
Congress on Sound and Vibration, Stockholm, Sweden, 7–10 July 2003.
3. Zou, C.; Wang, Y.; Moore, J.A.; Sanayei, M. Train-induced field vibration measurements of ground and
over-track buildings. Sci. Total Environ. 2017, 575, 1339–1351. [CrossRef] [PubMed]
4. López-Mendoza, D.; Romero, A.; Connolly, D.P.; Galvína, P. Scoping assessment of building vibration
induced by railway traffic. Soil Dyn. Earthq. Eng. 2017, 93, 147–161. [CrossRef]
5. BS 7385-2:1993. Evaluation and Measurement for Vibration in Buildings—Part 2: Guide to Damage Levels from
Ground Borne Vibration; British Standards Institution: London, UK, 1993.
6. Kouroussis, G.; Kinet, D.; Moeyaert, V.; Dupuy, J.; Caucheteur, C. Railway structure monitoring solutions
using fibre Bragg grating sensors. Int. J. Rail Transp. 2016, 4, 135–150. [CrossRef]
7. Filograno, M.L.; Corredera, P.; Rodríguez-Barrios, A.; Martín-López, S.; Rodríguez-Plaza, M.;
Andrés-Alguacil, A.; González-Herráez, M. Real Time monitoring of railway traffic using fiber Bragg
grating sensors. IEEE Sens. J. 2011, 12, 85–92. [CrossRef]
8. Paolozzi, A.; Paris, C.; Vendittozzi, C.; Felli, F.; Mongelli, M.; De Canio, G.; Colucci, A.; Asanuma, H. Test
of FBG sensors for monitoring high pressure pipes. Proc. SPIE Sens. Smart Struct. Technol. Civ. Mech.
Aerosp. Syst. 2017, 10168, 101681Q1-8. [CrossRef]
9. Huang, J.; Zhou, Z.; Zhang, L.; Chen, J.; Ji, C.; Pham, D.T. Strain modal analysis of small and light pipes
using distributed fibre Bragg grating sensors. Sensors 2016, 16, 1583. [CrossRef] [PubMed]
10. Salo, J.; Korhonen, I. Calculated estimate of FBG sensor’s suitability for beam vibration and strain measuring.
Measurement 2014, 47, 178–183. [CrossRef]
11. Bassam, A.; Ansari, F. A strain monitoring approach for estimating lateral displacements of concrete columns
under seismic excitations. J. Intell. Mater. Syst. Struct. 2016, 27, 71–80. [CrossRef]
12. Kang, M.K.; Park, D.J.; Lee, S.S. Strain measurements on a cantilever beam with fiber Bragg grating sensors
using a pair of collimators. Int. J. Precis. Eng. Manuf. 2012, 13, 455–458. [CrossRef]
13. Peeters, B.; Marques dos Santos, F.L.; Pereira, A.; Araujo, F. On the use of optical fiber Bragg grating (FBG)
sensor technology for strain modal analysis. AIP Conf. Proc. 2014, 1600, 39.
14. Young, W.C.; Budynas, R.G. Roark’s Formulas for Stress and Strain, 7th ed.; The McGraw-Hill Publishers:
New York, NY, USA, 2002.
15. Yao, K.; Lon, Q.; Jiang, Z.; Zhao, N.; Tian, B.; Shi, P.; Peng, G.D. Modeling and analysis of a combined
stress-vibration fiber Bragg grating sensor. Sensors 2018, 18, 743. [CrossRef] [PubMed]
16. Shabana, A.A. Chapter 4: Forced vibration. In Theory of Vibration: An Introduction; Springer: New York, NY,
USA, 1996. [CrossRef]
17. Singh, G.K.; Kazzaz, S.A.S. Induction machine drive condition monitoring and diagnostic research—A
survey. Electr. Power Syst. Res. 2003, 64, 145–158. [CrossRef]
18. Available online: http://www.brown.edu/Departments/Engineering/Courses/En4/Notes/vibrations_
forced/vibrations_forced.htm (accessed on 17 May 2018).

© 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).