Journal of

composites science

Article
Vibration Analysis of a Composite Concrete/GFRP
Slab Induced by Human Activities
Paulo Junges 1, *, Henriette Lebre La Rovere 2 and Roberto Caldas de Andrade Pinto 2
1 Civil Engineering Department, Federal University of Mato Grosso, Cuiabá 78060-900, Brazil
2 Civil Engineering Department, Federal University of Santa Catarina, Florianópolis 88040-900, Brazil;
henriettelarovere@gmail.com (H.L.L.R.); r.pinto@ufsc.br (R.C.d.A.P.)
* Correspondence: paulojunges@gmail.com; Tel.: +55-65-98116-5655

Received: 24 August 2017; Accepted: 11 September 2017; Published: 29 September 2017

Abstract: Fiber-reinforced polymer (FRP) materials have been introduced recently in the construction
of new structural systems, particularly in footbridge systems. Innovative systems that combine
concrete with FRP materials lead to lighter and more slender structures as compared to conventional
reinforced concrete structures, which can bring about vibration problems. In this work, a vibration
analysis of a composite slab subjected to human activities is performed, both experimentally and
numerically. The slab is composed of a concrete top laid on glass fiber-reinforced polymer (GFRP)
I-section pultruded profiles. In the experimental analysis, two prototypes of 0.80 m width and 4.00 m
span, representing a slab strip, were subjected to walking and jumping by several volunteers. In the
numerical analysis, the slab was modeled by finite elements under dynamic loadings that simulate
walking and jumping. Both the experimental and numerical results have indicated that the dynamic
behavior under human activities of the composite slab must be considered in the design.

Keywords: composite slab; GFRP profiles; vibrations

1. Introduction
Among the advanced composite materials, fiber-reinforced polymers (FRPs) stand out, due to
some favorable characteristics such as light weight, high specific strength and stiffness, as well as high
resistance to corrosion and fatigue. These characteristics encouraged the engineers to use FRP not just
for strengthening, but also to build new structures, particularly in the case of structures located near
maritime shores or those exposed to aggressive atmospheric agents. There are several examples of
application of FRP materials to footbridges, such as the Aberfeldy Footbridge in Scotland, and the one
over the rail system in Kolding, Denmark [1].
Starting from the 1990s, engineers and researchers awake to the advantages of combining concrete
and FRP materials. While concrete aids in compressive resistance and stability, FRPs provide tensile
resistance and can be used as permanent formwork to the fresh concrete. By combining these materials,
the overall cost of the innovative system is reduced as compared to systems built only with FRP.
Following this research line on innovative systems, a composite slab made of a fiber-reinforced
concrete top laid on glass fiber reinforced polymer (GFRP) I-section profiles is being developed at the
Federal University of Santa Catarina, Brazil (UFSC), for footbridge deck applications. The slab consists
of a concrete top laid on GFRP I-section pultruded profiles, filled in with foam blocks, as shown
in Figure 1. In previous works, this slab system was verified under static loads, both numerically
and experimentally [1,2], showing very good results under bending for two systems, with different
I-section profiles. Initial studies on the slab durability and its behavior under concentrated loads have
also been performed [3]. As this slab system is lighter and slenderer as compared to conventional
reinforced concrete slabs, it becomes necessary to verify its behavior under dynamic loading.

J. Compos. Sci. 2017, 1, 11; doi:10.3390/jcs1020011 www.mdpi.com/journal/jcs

J. Compos. Sci. 2017, 1, x FOR PEER REVIEW 2 of 20

compared
J. Compos. Sci.to conventional
2017, 1, 11 reinforced concrete slabs, it becomes necessary to verify its behavior under
2 of 20
dynamic loading.

Figure 1.
Figure 1. Composite
Composite slab
slab concrete/glass
concrete/glass fiber-reinforced
fiber-reinforced polymer
polymer (GFRP)
(GFRP) profiles.
profiles. Reproduced with
Reproduced with
permission from Santos Neto, A.B.S., published by Federal University of Santa Catarina, 2006.
permission from Santos Neto, A.B.S., published by Federal University of Santa Catarina, 2006.

The aim of this work is to investigate the dynamic behavior of the slab system under
The aim of this work is to investigate the dynamic behavior of the slab system under development
development at UFSC, regarding to vibrations induced by human activities such as walking and
at UFSC, regarding to vibrations induced by human activities such as walking and jumping. Initially,
jumping. Initially, a brief literature review on modeling of the dynamic loads caused by human
a brief literature review on modeling of the dynamic loads caused by human activities such as walking
activities such as walking and jumping is presented, describing the mathematical models used in the
and jumping is presented, describing the mathematical models used in the numerical analysis of
numerical analysis of the slab system. Next, the slab system and the material properties of its
the slab system. Next, the slab system and the material properties of its components are described.
components are described. The dynamic behavior of the slab system was verified by means of
The dynamic behavior of the slab system was verified by means of experimental and numerical
experimental and numerical analyses. In the experimental analysis, two prototypes of 0.80 m width
analyses. In the experimental analysis, two prototypes of 0.80 m width and 4.00 m span were subjected
and 4.00 m span were subjected to walking and jumping by several volunteers. In the numerical
to walking and jumping by several volunteers. In the numerical analysis, the slab prototype was
analysis, the slab prototype was modeled by finite elements, using SAP2000 software [4]. The slab
modeled by finite
was subjected to elements, using SAP2000
dynamic loads software
representing [4]. The slab was
the experimental subjected
testing, applyingto dynamic loads
mathematical
representing the experimental testing, applying mathematical models developed by other
models developed by other authors. The elastic properties of concrete were estimated using equations authors.
The
givenelastic properties
in design codes,of concretethe
whereas were estimated
elastic using
properties of equations
the GFRP given in were
profiles design codes, whereas
obtained the
analytically
elastic
and alsoproperties of the GFRP profiles were obtained analytically and also experimentally.
experimentally.
Results
Results of the experimental
of the experimental andand ofof the
the numerical
numerical analysis
analysis showed
showed that
that even
even though
though the
the
fundamental frequency of the slab under study is much larger than the limit
fundamental frequency of the slab under study is much larger than the limit recommended by recommended by
international codes regarding
international codes regardinghuman
humancomfort
comfortcriteria,
criteria,a adynamic
dynamicanalysis
analysisofof
thethe slab
slab still
still needs
needs to
to be
be performed.
performed.
2. Human Sensitivity to Vibrations—Brief Literature Review
2. Human Sensitivity to Vibrations—Brief Literature Review
The development of structural systems as well as of advanced materials in the past decades gave
The development of structural systems as well as of advanced materials in the past decades gave
rise to the construction of lighter and more slender structures with larger spans. Such structures are
rise to the construction of lighter and more slender structures with larger spans. Such structures are
in general more susceptible to vibration effects caused by dynamic actions, such as wind, traffic of
in general more susceptible to vibration effects caused by dynamic actions, such as wind, traffic of
vehicles and people, earthquakes, as well as equipment in nearby construction sites.
vehicles and people, earthquakes, as well as equipment in nearby construction sites.
People are daily exposed to vibrations on floor slabs and on footbridges, caused by different
People are daily exposed to vibrations on floor slabs and on footbridges, caused by different
sources of excitation. The evaluation of human sensitivity to these vibrations involves psychological
sources of excitation. The evaluation of human sensitivity to these vibrations involves psychological
and physical aspects [5]. As for footbridges and walkways, the most import dynamic actions are the
and physical aspects [5]. As for footbridges and walkways, the most import dynamic actions are the
ones related to human activities like walking and jumping. The resultant human-induced vibrations
ones related to human activities like walking and jumping. The resultant human-induced vibrations
may cause serviceability problems as well as discomfort to the users [6]. The main factors that
may cause serviceability problems as well as discomfort to the users [6]. The main factors that may
may influence human sensitivity are the position (standing, sitting or lying), the type of activity
influence human sensitivity are the position (standing, sitting or lying), the type of activity being
being performed, age, gender, mood, vibration frequency, displacement amplitudes, damping, and
performed, age, gender, mood, vibration frequency, displacement amplitudes, damping, and
acceleration of the dynamic excitation.
acceleration of the dynamic excitation.
Reiher and Meister [7] investigated the effect of harmonic vibrations on people in different
Reiher and Meister [7] investigated the effect of harmonic vibrations on people in different
postures on a test platform, subjected to various excitation frequencies, amplitudes and directions.
postures on a test platform, subjected to various excitation frequencies, amplitudes and directions.
Results from this study were presented in a scale graph of human perception as a function of excitation
J. Compos.
J. Compos. Sci.
Sci. 2017,
2017, 1,
1, x
11FOR PEER REVIEW 33 of
of 20
20

Results from this study were presented in a scale graph of human perception as a function of
excitation frequency
frequency and amplitude.
and amplitude. Similar
Similar scales basedscales based on experiments
on experiments of vertical of vertical short-duration
short-duration vibration,
vibration,
like like
the one the one experienced
experienced by footbridge
by footbridge users, have users,
also have
been also been proposed.
proposed. Figure 2 the
Figure 2 presents presents
scale
the scale
graph graph by
proposed proposed
Goldmanby[8].Goldman
This scale[8]. Thisanscale
shows shows between
interaction an interaction
the peakbetween the peak
acceleration and
acceleration
the frequencyand the frequency
to human to human perception.
perception.

Figure 2. Scale used to measure human sensitivity to vibrations suggested

suggested by
by [8].
[8].

Other design
Other designguidelines,
guidelines, however,
however, do not
do use
notpeak
useacceleration, but rather,
peak acceleration, butapply the root-mean-
rather, apply the
root-mean-square (RMS) acceleration to verify the level of admissible vibrations. EN 1990basis
square (RMS) acceleration to verify the level of admissible vibrations. EN 1990 [9], the Eurocode [9],
of structural
the Eurocode design,
basis of allows a maximum
structural RMS aacceleration
design, allows maximum RMS of 0.7acceleration
m/s2, and of a smallest
0.7 m/s2 ,natural
and a
frequency
smallest of 5 Hz
natural for comfort
frequency of 5 Hzcriteria. ISO 10137
for comfort [10]ISO
criteria. guidelines forguidelines
10137 [10] serviceability in buildings
for serviceability
suggests using the curves presented in Figure 3. This guideline permits higher RMS
in buildings suggests using the curves presented in Figure 3. This guideline permits higher RMS accelerations
when a general
accelerations whenassessment of footbridge
a general assessment vibrations vibrations
of footbridge is performed than to the
is performed scenario
than of a person
to the scenario of a
standing at midspan while another crosses the footbridge.
person standing at midspan while another crosses the footbridge.

Dynamic Loads Generated

Generated by
by Human
Human Activities
Activities
dynamic loads
The dynamic loadsgenerated
generatedby byhuman
humanactivities
activitiescan
canbebeseparated
separated into
into twotwo groups,
groups, according
according to
to the person-structure interaction: (a) when there is loss of contact with the
the person-structure interaction: (a) when there is loss of contact with the structure, like in running structure, like in running
and jumping; (b) when when there
there is
is not
not aa loss
loss ofof contact
contact withwith the
the structure,
structure, likelike in
in walking.
walking.
While walking or running, a human being produces a dynamic loading that cancan
While walking or running, a human being produces a dynamic loading that be be represented
represented by
abyforce
a force varying
varying in time
in time with components
with components in three in directions:
three directions:
vertical,vertical,
lateral andlateral and frontal.
frontal. The
The vertical
vertical component
component has been has been investigated
the most the most investigated
one, since itone, since greater
presents it presents greaterasamplitudes
amplitudes compared to as
compared to the other components [5]. This dynamic force is produced
the other components [5]. This dynamic force is produced by the acceleration and deceleration of the by the acceleration and
deceleration
body mass. For of the body mass.
walking, it has For
beenwalking,
stated that it has
its been statedrange
frequency that its frequency
is between 1.6range
to 2.4isHz
between
[11]. 1.6
to 2.4Even
Hz [11].
though there has not been the same comprehensive investigation for running or jumping
Evenitthough
activities, there has
is considered thatnot
thebeen
typicalthefrequency
same comprehensive
range for running investigation
is 2.0 to 3.5 forHzrunning
while foror jumping
activities,
is 1.8 to 3.4itHzis [11].
considered that the typical frequency range for running is 2.0 to 3.5 Hz while for
jumping is 1.8 to 3.4 Hz [11].

Mathematical Modeling for Walking and Jumping

 To describe loads induced by walking and jumping, mathematical models in the frequency
domain as well as in the time domain can be found in the literature. In this work, only a deterministic
time domain model that aims to establish a general load model for each type of human activity is
considered. It is assumed that the induced force is equal in both feet. This force can be represented
J. Compos. Sci. 2017, 1, 11 4 of 20
by a Fourier series as described by Equation (1):

Figure 3. Curves for level of vibrations in the vertical direction for walkways of [10].
Figure 3. Curves for level of vibrations in the vertical direction for walkways of [10].

Mathematical Modeling for Walking and Jumping

FP t = G + ∑ni=1 Gαi sen 2πifP t - ϕi , (1)
To describe loads induced by walking and jumping, mathematical models in the frequency
domain as well as in the time domain can be found in the literature. In this work, only a deterministic
where is the resulting transient load; is the time in seconds; represents the person’s
time domain model that aims to establish a general load model for each type of human activity is
weight; is the number of the ith harmonic; is the number of harmonics to be considered;
considered. It is assumed that the induced force is equal in both feet. This force can be represented by
represents the dynamic coefficient for the ith harmonic; is the step frequency and is the phase
a Fourier series as described by Equation (1):
angle, between the ith harmonic and the first one.
For walking, Varella [12] proposed a modification of Equation (1), as presented in Equations (2)
FP (t) = G + ∑ in=1 Gαi sen(2πif P t − ϕi ), (1)
to (5). He considered that the human induced loads can be appropriately described by considering
the firstF four harmonics of the Fourier series. This load model considers that the generated time
where P ( t ) is the resulting transient load; t is the time in seconds; G represents the person’s weight; i
function has
is the number of a space
the and time description.
ith harmonic; This more
n is the number realistic load
of harmonics to model has beenαused
be considered; in similar
i represents the
studies
dynamic [13,14], and therefore
coefficient for the ithisharmonic;
adopted here.
f is the step frequency and ϕ is the phase angle, between
P i
the ith harmonic and the first one.
For walking, Varella [12] proposed a modification of Equation (1), as presented in Equations (2)
to (5). He considered that the human induced loads can be appropriately described by considering
the first four harmonics of the Fourier series. This load model considers that the generated time
J. Compos. Sci. 2017, 1, 11 5 of 20

function has a space and time description. This more realistic load model has been used in similar
studies [13,14], and therefore is adopted here.

  
f mi Fm − G
 t + G for 0 ≤ t < 0.04TP
h 0.04TP


 i
 C1 (t − 0.04TP )



 f mi Fm 0.02 TP + 1 for 0.04TP ≤ t < 0.06TP

FP (t) = Fm for 0.06TP ≤ t < 0.15TP (2)
 n
G + ∑ Gαi sen[2πif P (t + 0.1TP ) − ϕi ] for 0.15TP ≤ t < 0.90TP






 i =1  
10( G − C2 ) TtP − 1 + G for 0.90TP ≤ t < TP



!
n
Fm = G 1 + ∑ αi (3)
i =1
 
1
C1 = − 1 (4)
f mi
(
G (1 − α2 ) if n = 3
C2 = (5)
G (1 − α2 + α4 ) if n = 4

In Equation (2), Fm is the maximum value of the Fourier series, which is given by Equation (3);
fmi is the amplification factor due to heel impact; Tp is the step period; and C1 and C2 are coefficients
defined by Equations (4) and (5), respectively.
A value of 1.12 may be used for the amplification factor due to heel impact even though this value
could vary according to the walking person [13]. For the dynamic coefficients αi , Varela [12] presented
the best-fit polynomial functions (Equations (6) to (9)) obtained from the data presented by Rainer,
Pernica and Allen [15]:

α1 = − 0.22169 f P 3 + 1.11946 f P 2 − 1.44748 f P + 0.5967 (6)

α2 = − 0.012037 (2 f P )3 + 0.1494 (2 f P )2 − 0.53146 (2 f P ) + 0.6285 (7)

α3 = 0.00009068 (3 f P )5 − 0.0021066 (3 f P )4 + 0.018364 (3 f P )3 − 0.077278 (3 f P )2 + 0.17593 (3 f P ) − 0.1477 (8)

α4 = − 0.00051715 (4 f P )4 − 0.014388 (4 f P )3 + 0.14562 (4 f P )2 − 0.62994 (4 f P ) − 1.018469 (9)

The phase angles are equal to: ϕ1 = 0; ϕ2 = π/2; ϕ3 = π; ϕ4 = 3π/2. By using this model, the loads
induced by walking observed by Ohlsson [16] could be appropriately described, as can be seen in
Figure 4. In this work, Equations (2) to (9) will be used to model the walking load in the numerical
analysis (subsection 4.2).
David et al. [17] modeled the force generated by a vigorous jumping action, based on
Bachmann et al. [11], by considering a frequency of jumping of 3 Hz, and the first three harmonics
with dynamic coefficients of 1.7, 1.1, and 0.5, respectively. The phase angles were ϕ1 = 0; ϕ2 = ϕ3 =
π [1 − ( f p tc )] = 1, with the contact time with the structure, tc as 0.2 s. Figure 5 presents the resultant
load function proposed, Fi (t). The jumping transient load can then be obtained by multiplying Fi (t) by
the person’s weight, G. This resulting load will be used in the numerical analysis (subsection 4.3) to
model the jumping cases.
J. Compos. Sci. 2017, 1, 11 6 of 20
J. Compos. Sci. 2017, 1, x FOR PEER REVIEW 6 of 20
J. Compos. Sci. 2017, 1, x FOR PEER REVIEW 6 of 20

Figure 4.
Figure 4. Load
Loadfunction
functioninduced
inducedbybywalking
walkingproposed
proposedbyby[12], compared
[12], compared to to
results obtained
results by
obtained
Figure
by 4. Load function
[16].Reproduced
[16].Reproducedwith
with induced by
permission
permission walking
from
from proposed
Varela,
Varela,W.D., by [12], compared
W.D.,published
published by FederaltoUniversity
by Federal results obtained
University by
of Rio de
de
[16].Reproduced
Janeiro,
Janeiro, 2004.
2004. with permission from Varela, W.D., published by Federal University of Rio de
Janeiro, 2004.

Figure 5. Load function induced by jumping used by [17].

Figure5.
Figure 5. Load
Load function
function induced
induced by
by jumping
jumpingused
usedby
by[17].
[17].
3. Composite Slab Concrete/GFRP Profiles
3. Composite
3. Composite SlabSlab Concrete/GFRP
Concrete/GFRP Profiles
Profiles
The slab system used in this study is shown schematically in Figure 6. Narrow flanged (NF) I-
TheGFRP
section
The slab system
slab system used
profilesused
within this
in the study
study isis shown
this dimensions shownshownschematically
in Figure in
schematically inFigure
6a have 6.
Figure 6.Narrow
been flanged
selected.
Narrow The(NF)
flanged foam
(NF) I-
section GFRP
(expanded
I-section GFRP profiles
profiles with
polystyrene or the
EPS)
with dimensions
theblocks used for
dimensions shown inare
fillingin
shown Figure 6a have
have been
non-structural
Figure 6a been selected.
and selected.
have The
beenThe foam
cut foam
for a
(expanded
perfect fit; polystyrene
they have theor EPS)
usual blocks
dimensions used for
employedfilling
in are non-structural
precast concrete and
slabs. have
The
(expanded polystyrene or EPS) blocks used for filling are non-structural and have been cut for a perfect been
profiles cut
and for
thea
perfect
foam
fit; they fit;
blocks
havethey
thehave
serve also
usualthe
asusual dimensions
formwork
dimensions for theemployed
employed wetinconcrete.in precast
precast concrete
They are
concrete slabs.
designed
slabs. The to The profiles
sustain
profiles and
constructive
and the the
foam
foam
loads, blocks
avoidingserve also
thus theas formwork
use of for
bracing. the wet concrete. They are designed to
blocks serve also as formwork for the wet concrete. They are designed to sustain constructive loads, sustain constructive
loads,Inavoiding
avoiding order thethus
thus to usethe
restrain use of bracing.
cracking
of bracing. due to plastic shrinkage, short polypropylene fibers are added to
In ordertop,
the concrete to restrain crackingfraction
with a volume due toof plastic shrinkage,
0.1% [1,18]. short polypropylene
An epoxy-based fibers are
resin is utilized for added
bonding to
the concrete top,
to thewith a volume
profiles fraction
at the of 0.1%
interface. This[1,18].
resinAn epoxy-based
avoids resin is utilized
water penetration, for bonding
and thus aids to
the concrete
prevent alkalitoattack
the profiles at theglass
in the profile interface.
fibers.This resin avoids water penetration, and thus aids to
prevent alkali attack in the profile glass fibers.
 After curing of concrete, the two components, concrete and GFRP profiles, are joined together
and behave structurally as a precast slab. A spacing of 400 mm was defined between the profile axes,
while the concrete top thickness was chosen to be 40 mm, based on the recommendations for precast
slabs given in the Brazilian Code NBR 14859-1 [19]. These dimensions avoid buckling of the profile
J. Compos. Sci. 2017, 1, 11 7 of 20
walls and shear failure of the concrete top [20].

(a) (b)

Figure6.
Figure 6. (a)
(a) Dimension
Dimension of
of I-section
I-section GFRP
GFRP profile;
profile; (b)
(b) Cross-section
Cross-section of
of the
the representative
representativestrip
strip utilized
utilized
to study the composite slab behavior.
to study the composite slab behavior.

A one-way action was assumed for the flexure design of the composite slab system. A strip of
In order to restrain cracking due to plastic shrinkage, short polypropylene fibers are added to the
0.80m width was taken as a representative section, as displayed in Figure 6b. In the design, care was
concrete top, with a volume fraction of 0.1% [1,18]. An epoxy-based resin is utilized for bonding the
taken to keep the profiles working in tension and the concrete top almost entirely in compression,
concrete to the profiles at the interface. This resin avoids water penetration, and thus aids to prevent
taking at the most advantage of both materials.
alkali attack in the profile glass fibers.
This slab system can be employed either supported by longitudinal beams, with the profiles axis
After curing of concrete, the two components, concrete and GFRP profiles, are joined together
disposed perpendicular to the footbridge longitudinal axis, or supported by transverse beams, with
and behave structurally as a precast slab. A spacing of 400 mm was defined between the profile axes,
the profiles axis disposed parallel to the footbridge longitudinal axis.
while the concrete top thickness was chosen to be 40 mm, based on the recommendations for precast
slabs given inProperties
3.1. Material the Brazilian Code NBR 14859-1 [19]. These dimensions avoid buckling of the profile
walls and shear failure of the concrete top [20].
AThe selectedaction
one-way GFRPwasprofiles are composed
assumed of fiber
for the flexure glassofrovings
design (disposed
the composite parallel
slab to the
system. profile
A strip of
longitudinal axis) embedded in a polyester matrix, with a fiber volume fraction
0.80m width was taken as a representative section, as displayed in Figure 6b. In the design, care was of 60%, and of
laminates
taken to keepmadetheofprofiles
continuous strand
working in mats (CSM),
tension and consisting
the concreteof top
randomly-oriented
almost entirely fibers.
in compression,
It is initially assumed that the profile
taking at the most advantage of both materials. web and flanges are laminated composites with the same
lay-up and equivalent orthotropic mechanical properties. The mechanical elastic
This slab system can be employed either supported by longitudinal beams, with the profiles properties of each
axis
individual
disposed laminae can to
perpendicular bethe
obtained using
footbridge the Rule ofaxis,
longitudinal Mixtures by using
or supported bythe elastic properties
transverse beams, withof the
the
fibers, resin, and mats as given by the fabricator [21].
profiles axis disposed parallel to the footbridge longitudinal axis. The profile elastic properties can then be
estimated from the laminae properties by using the Classical Lamination Theory with results shown
3.1. Material
in Table 1. Properties
The selected GFRP profiles are composed of fiber glass rovings (disposed parallel to the profile
Table 1. Elastic properties of the laminates that form the GFRP profile web and flanges (considered
longitudinal axis) embedded in a polyester matrix, with a fiber volume fraction of 60%, and of laminates
as an equivalent orthotropic material)a.
made of continuous strand mats (CSM), consisting of randomly-oriented fibers.
Elastic Properties
It is initially assumed that the profile E1 (GPa) E2 (GPa)
web and flanges are laminatedG12 composites
(GPa) ν12 (GPa)
with the same
lay-up and equivalent orthotropic
Laminates of GFRP profiles mechanical properties. The mechanical elastic properties of each
individual laminae can be obtained 26.73
using the Rule of 7.19
Mixtures by using2.44the elastic 0.341
properties of
(I section)
the fibers, resin, and mats as given by the fabricator [21]. The profile elastic properties can then be
a where 1 is the longitudinal direction, parallel to the longitudinal profile axis, and 2 is the transversal
estimated from the laminae properties by using the Classical Lamination Theory with results shown in
direction of the profile web and flanges.
Table 1.
The elastic properties of the GFRP profiles were also obtained experimentally from three-point
1. Elastic
Tabletests
bending properties of
by following a the laminatesproposed
procedure that form the
byGFRP
Bankprofile web and
[22] based onflanges (considered asBeam
the Timoshenko
an equivalent orthotropic material) a .
Theory. It considers that the profile behaves as an equivalent isotropic material. The equivalent
longitudinal (E) and shear (G) moduli of the GFRP profiles were found to be 26.21 and 4.29 GPa,
Elastic Properties E1 (GPa) E2 (GPa) G12 (GPa) ν12 (GPa)
respectively. A more detailed description of this procedure and experiments can be found in Santos
Laminates
Neto and La Rovere [23]. of GFRP profiles
26.73 7.19 2.44 0.341
(I section)
a where 1 is the longitudinal direction, parallel to the longitudinal profile axis, and 2 is the transversal direction of
the profile web and flanges.
J. Compos. Sci. 2017, 1, 11 8 of 20

The elastic properties of the GFRP profiles were also obtained experimentally from three-point
bending tests by following a procedure proposed by Bank [22] based on the Timoshenko Beam Theory.
It considers that the profile behaves as an equivalent isotropic material. The equivalent longitudinal
(E) and shear (G) moduli of the GFRP profiles were found to be 26.21 and 4.29 GPa, respectively.
A more detailed description of this procedure and experiments can be found in Santos Neto and La
Rovere [23].
Concrete was designed for a compressive strength of 30 MPa at 28 days. The secant modulus of
elasticity (Ec ) was obtained according to ACI 318 recommendation [24] yielding a value of 26.07 GPa.
The usual value of 0.2 was adopted for Poisson´s coefficient, resulting in a shear modulus (Gc ) of
10.86 GPa.

3.2. Design and Static Analysis

Previous studies [1] have shown that the Serviceability Limit State (SLS) tends to govern the
design of composite slabs made of concrete and GFRP materials for footbridge deck applications. The
composite slab under study was then initially designed under bending for service loads, without
considering long term effects. A representative slab strip of 0.8 m width (see section shown in Figure 6)
was selected for the analysis. A total dead load (DL) equal to the self-weight (0.846 kN/m) plus an
additional load of 1.2 kN/m, due to wearing surface and guardrails, plus a live load (LL) of 4 kN/m
corresponding to the usual pedestrian load of 5 kN/m2 , were considered acting along the span of
the simply supported slab strip. One-way action was assumed and the deflections in the slab were
calculated using the Timoshenko Beam Theory (TBT), combined with the Transformed Area Approach.
The maximum deflection in the slab, at midspan, considering shear deformation, can be found by
Equation (10):
5qL4 qL2 f s
v = + (10)
384 EI 8GA
where: v is the maximum displacement; L is the span; q is the design service load uniformly distributed
along the span; EI is the flexural stiffness of the transformed section; GA/fs is the shear stiffness of the
transformed section, in which fs is the shear factor.
The design service load was found considering the quasi-permanent load combination
(q = DL + 0.4LL) [9]. The slab section shown in Figure 6 was transformed into an equivalent GFRP
section, by assuming that concrete and GFRP were perfectly bonded and behaved linear-elastically.
It was also assumed that concrete remains uncracked under service loads. By using the equivalent
isotropic moduli of the GFRP profiles E, G and the estimated values for the concrete moduli Ec , Gc ,
defined in Section 3.1, the resulting properties of the transformed section were calculated in Table 2.

Table 2. Transformed section properties b .

x (mm) EI (kNm2 ) GĀ (kN) n

30.04 1.263 × 103 1.376 × 104 0.995
b where x is the neutral axis depth; EI is the flexural stiffness; GĀ is the shear stiffness, in which Ā = A/fs and fs is
the shear factor; n = Ec /EGFRP is the moduli ratio.

By inserting all of these values in Equation (10), and by adopting the allowable limit of L/250
given in the Brazilian code NBR 6118 [15] recommendations, a maximum allowable span of 4.65 m
was found for the composite slab strip. It then verified the assumption that the concrete top does not
crack under service loads.
Next, the composite slab was verified at the Ultimate Limit State (ULS). Three possible failure
modes were considered: (1) flexural failure caused by crushing of the extreme compression fiber in
the concrete top; (2) shear failure in the GFRP profile at the top flange/web intersection, and (3) bond
failure at the GFRP profiles/concrete top interface. For flexural failure, the ultimate bending moment
J. Compos. Sci. 2017, 1, 11 9 of 20

Mu can be calculated from force equilibrium, strain-compatibility conditions and the constitutive
behavior of the materials. It was assumed that GFRP materials behave linear-elastically and that
concrete was a nonlinear elastic material. The concrete top and the GFRP profiles were assumed to be
perfectly bonded and the tensile strength of concrete was neglected. The equivalent rectangular stress
block as defined by the ACI 318 [24] was adopted. Results of ultimate efforts are shown in Table 3.
In this table, the ultimate shear Vu1 was obtained using an estimated theoretical value for the ultimate
shear stress in the GFRP laminae, obtained from Halphin-Tsai equations [21], and Vu2 was found from
the bond strength at the interface concrete/GFRP, which has been measured experimentally [1].

Table 3. Ultimate Limit State (ULS): Design and ultimate efforts for the composite slab under uniform
load (L = 4.65 m).

Effort (a) Bending (b) Shear Web/Flange (c) Bond Shear (Interface)
Design Effort Md = 23.93 kN·m V d1 = 20.60 kN V d2 = 20.60 kN
Ultimate Effort Mu = 92.26 kN·m V u1 = 40.10 kN V u2 = 58.02 kN
Safety Factor Φb = 0.26 Φs1 = 0.51 Φs2 = 0.36

For the maximum span of 4.65 m, the design efforts in the simply supported slab strip were found
by considering the load combination 1.2DL + 1.6LL [24]. These efforts are shown in Table 3, with the
resulting safety factors (design/ultimate efforts). As it can be observed from Table 3, the critical failure
mode is shear failure in the GFRP profile at the top flange/web intersection (safety factor Φs1 = 0.51).
Although the required safety factors have not yet been established in the Design Codes for the case of
concrete/GFRP composite slabs, a safety factor of at most 0.50 should be adopted, since a brittle kind
of failure is expected in such structures. Hence, in order to attend both serviceability and ultimate
limit states, the span of the composite slab was reduced to 4.0 m (Φs1 = 0.44).
In order to verify the stiffness and the ultimate efforts estimated in the design of the composite
slab under static bending, experimental tests were also carried out. Two slab prototypes of 0.8 m width
and 4.0 m of span were tested under 4-point test bending. The loads (P) were applied at a distance (a)
of 1.575 m from the supports.
Initially the slab stiffness was investigated by applying on the prototypes two concentrated loads
(P) of 4.63 kN, in such a way to obtain the same maximum bending moment produced by a uniformly
distributed load (3.646 kN/m) for the service load combination (DL + 0.4 LL). It was observed during
the tests that the slab behavior was basically linear elastic under service.
The maximum displacement at midspan can be obtained theoretically using TBT, by means of
Equation (11):  
Pa 3L2 − 4a2 Paf s
v = + (11)
24 EI GA
where: v is the maximum displacement; L is the span; P are the concentrated loads applied
symmetrically on the slab; a is the distance between the applied loads and the supports; EI is the
flexural stiffness and GA/fs is the shear stiffness of the transformed section (see Table 2).
Comparison between theoretical and experimental values of maximum displacement is shown
in Table 4. It can be observed from the table that the theoretical value gives a good estimate for the
average value obtained experimentally from the slab prototypes.

Table 4. Serviceability Limit State (SLS): Comparison of maximum displacement obtained theoretically
and experimentally for the composite slab under 4-point bending (L = 4.00 m; P = 4.63 kN).

Theoretical (TBT) Experimental Limit (L/250)

Prototype I = 10.00 mm
9.69 mm 16.00 mm
Prototype II = 7.92 mm
- Average value = 8.96 mm -
J. Compos. Sci. 2017, 1, 11 10 of 20

Next, the loads were increased on the prototypes (I and II) until failure. Comparison between
ultimate efforts estimated theoretically and obtained experimentally is summarized in Table 5. It can
be observed that the prototypes showed different failure modes (I showed bond shear failure, and
II showed shear failure in the profile at the web/flange interface). In prototype I, a premature bond
shear failure occurred at the interface of the concrete top and only one of the GFRP profiles, which
was probably due to bad finishing of the concrete top observed in a small region, giving rise to a
non-uniform load application along the slab width. Nevertheless, in spite of the different failure
modes, both prototypes fail at a load (V I = 38.5 kN and V II = 39.0 kN), close to the estimated one
(V u = 40.1 kN).
In the following, the dynamic behavior of the composite slab will be investigated.

Table 5. ULS: Comparison between ultimate efforts obtained theoretically and experimentally for the
composite slab under 4-point bending (L = 4.00 m).

(c) Bond Shear

Effort (a) Bending (b) Shear Web/Flange
(Interface)
Theoretical Ultimate effort Mu = 92.26 kN.m V u1 = 40.10 kN V u2 = 58.02 kN
Experimental Ultimate - - V I = 38.50 kN
Efforta - V II = 39.00 kNc -
c I and II refer to Prototype I and II, respectively.

4. Dynamic Behavior of the Composite Slab

The dynamic behavior of the composite slab under human activities was investigated both
experimentally and numerically. For the experimental analysis, two prototypes of 0.8 m width and
4.0 m of span, representing a slab strip (same geometry as used for the static tests), have been built and
subjected to walking and jumping by several volunteers. In the numerical analysis, the slab prototype
was modeled by finite elements and subjected to transient loads that simulate the tests, using the
mathematical models already described in Section 2.

4.1. Theoretical Fundamental Frequency of the Composite Slab

In order to aid planning the dynamic experimental program, the fundamental frequency of the
composite slab was initially estimated by a simplified theoretical model based on beam analogy
combined with the Transformed Section approach. The composite slab is considered as a linear
element (beam) of equivalent homogenous and isotropic material, with distributed mass along its
length. Thus, the beam equivalent fundamental frequency (ω 1 ) can be obtained from Equation (12),
using the equation based on Fourier series [25].
r
2 EI
ω1 = π , (12)
mL4

where: EI is the flexural stiffness of the transformed or homogenized section; L is the beam length; and
m is the beam mass per unit length.
By introducing the flexural stiffness of the homogenized section given in Table 2, the mass per
unit length (0.0863 ton/m) is found from the slab total self-weight, and the length L = 4.00 m in
Equation (12), the value of 74.66 rad/s is found for the fundamental angular frequency (ω 1 ), which
corresponds to a frequency f 1 of 11.88 Hz.

4.2. Experimental Analysis of the Composite Slab

Three types of vibration tests were performed on the slab prototypes: (i) heel-drop, (ii) walking
and (iii) jumping. The test setup and instrumentation utilized are illustrated in Figure 7. Vertical
J. Compos. Sci. 2017, 1, 11 11 of 20

accelerations were measured at midspan by means of two accelerometers, placed at the bottom of the
GFRP profiles, connected to a data acquisition system. Data were acquired at a time interval of 0.01s,
thus, frequencies up to 50 Hz could have been recorded. A total of five volunteers, whose characteristic
J. Compos. Sci. 2017, 1, x FOR PEER REVIEW 11 of 20
data are given in Table 6, participated in the tests. All volunteers used soft sole shoes during the tests.

Figure 7. Test set-up and

and instrumentation
instrumentation for
for the
the dynamic
dynamic tests
tests on
on the
the composite
composite slab.
slab.

Table 6.
Table 6. Characteristics
Characteristics of
of the
the volunteers
volunteers that
that participated
participated in
in the
the tests.
tests.

Volunteer Mass (kg) Height (m) Gender

Volunteer Mass (kg) Height (m) Gender
1 70 1.83 M
1 70 1.83 M
2 67 1.70 M
2 67 1.70 M
3 3 83 83 1.92
1.92 MM
4 4 53 53 1.63
1.63 FF
5 5 66 66 1.72
1.72 MM

Initially, the
Initially, the fundamental
fundamental frequency
frequency and
and the
the damping
damping factor
factor ofof the
the composite
composite slab
slab were
were
experimentally obtained through the heel-drop test. The average value obtained for
experimentally obtained through the heel-drop test. The average value obtained for the dampingthe damping
factor was
factor was utilized
utilized later
later in
in the
the numerical
numerical analysis
analysis (see
(see subsection
subsection 4.3).
4.3). In
In the
the sequence,
sequence, the
the composite
composite
slab was submitted to walking and jumping
slab was submitted to walking and jumping tests.tests.

4.2.1. Heel-Drop Test

4.2.1. Test
To obtain
To obtain the
the fundamental
fundamental frequency
frequency ofof the
the structure,
structure, the
the prototypes
prototypes of of the
the experimental
experimental
program were
program were initially
initially subjected
subjected to
to the
the heel-drop
heel-drop test,
test, as
as recommended
recommended by by the
the Canadian
Canadian Code
Code
CAN3-S16.1-M84 [26]. The
CAN3-S16.1-M84 Thetest
testconsisted
consistedofofa aperson
personstanding
standing onon
hishis
toes and
toes letting
and lettinghimself fallfall
himself on
his heels on the slab. Figure 8 presents the time-domain response as
on his heels on the slab. Figure 8 presents the time-domain response as well as well as the amplitude peak
peak
response of the slab
response slab in
inthe
thefrequency-domain
frequency-domainfor forone
oneprototype,
prototype, obtained
obtained from
fromthethe
testtest
performed
performedby
Volunteer
by 4. 4.
Volunteer Similar
Similargraphs
graphsnot shown
not shownhere
herewere
wereobtained
obtainedfrom
fromthe
thetests
tests performed
performed by the other
volunteers. The damping factor was found by means of the logarithmic decrement method in the
time-domain response.
J. Compos. Sci. 2017, 1, 11 12 of 20

volunteers. The damping factor was found by means of the logarithmic decrement method in the
time-domain
J. response.
Compos. Sci. 2017, 1, x FOR PEER REVIEW 12 of 20

Figure 8. Response
ResponseofofPrototype
PrototypeII IIto to
heel-drop performed
heel-drop by by
performed Volunteer 4 in4 terms
Volunteer of graphs:
in terms (a)
of graphs:
acceleration along time; (b) respective spectrum.
(a) acceleration along time; (b) respective spectrum.

Table 7 presents the fundamental frequency and the damping factor obtained for each prototype,
Table 7 presents the fundamental frequency and the damping factor obtained for each prototype,
from each volunteer test. It can be seen that the results in terms of fundamental frequency are quite
from each volunteer test. It can be seen that the results in terms of fundamental frequency are
uniform, whereas in terms of damping factor some variation is observed among the different
quite uniform, whereas in terms of damping factor some variation is observed among the different
volunteers. Average values of 12.64 Hz and 4.96% were obtained for the fundamental frequency and
volunteers. Average values of 12.64 Hz and 4.96% were obtained for the fundamental frequency and
the damping factor, respectively. This value of damping factor falls between the interval (1.5% to
the damping factor, respectively. This value of damping factor falls between the interval (1.5% to 5.0%)
5.0%) given in ISO 10137 [10] for fully composite concrete/steel beams with shear connectors, which
given in ISO 10137 [10] for fully composite concrete/steel beams with shear connectors, which is a
is a similar structure to the composite slab studied in this work.
similar structure to the composite slab studied in this work.
Table 7. Fundamental frequency and damping factor from heel-drop test.
Table 7. Fundamental frequency and damping factor from heel-drop test.
Natural Frequency (Hz) Damping Factor (%)
Volunteer Natural Frequency (Hz) Damping Factor (%)
Volunteer Prototype I Prototype
Prototype
Prototype I II II Prototype
Prototype II Prototype II II
Prototype
1 1 11.61 11.61 12.86
12.86 4.50
4.50 2.882.88
2 2 13.03 13.03 12.79
12.79 4.81
4.81 6.116.11
3 12.86 13.02 6.31 4.85
4 3 12.30 12.86 13.02
12.93 6.31
4.48 4.855.44
5 4 12.63 12.30 12.36
12.93 3.77
4.48 5.446.42
Mean value 12.64 Hz 4.96%
5 12.63 12.36 3.77 6.42

4.2.2. WalkingMean
Test value 12.64 Hz 4.96%

4.2.2.The walking
Walking Testtest consisted of a volunteer walking on the slab prototype, as shown in Figure 9,
resulting in an excitation frequency of about 2 Hz. To allow an interaction person-structure, i.e., that
The walking
the volunteer getstest consisted
in rhythm of the
with a volunteer walking
slab, a total on walks
of eight the slab
onprototype, as shown
each prototype werein Figure 9,
performed.
resulting
Results ofinthe
anslab
excitation frequency
response of about
to the walking 2 Hz.
tests areTo allow
given in an interaction
terms person-structure,
of graphs i.e., that
of acceleration (average
the volunteer gets in rhythm with the slab, a total of eight walks on each prototype
value obtained from the measurements of the two accelerometers) versus time, with its respective were performed.
Results
spectrum.of the slab10response
Figure to results
shows the the walking testsfrom
obtained are given inperformed
the test terms of graphs of acceleration
by Volunteer 4. Similar (average
results,
value obtained from the measurements of the two accelerometers) versus time, with
not shown here, were obtained from the tests performed by the other volunteers. Table 8 presents the its respective
spectrum. Figure 10 shows
high peak accelerations theRMS
and the results obtained obtained
accelerations from thefromtest the
performed by Volunteer
tests performed 4. Similar
by all volunteers.
results, not shown here, were obtained from the tests performed by the other volunteers. Table 8
presents the high peak accelerations and the RMS accelerations obtained from the tests performed by
all volunteers.
J. Compos. Sci. 2017, 1, 11 13 of 20
J. Compos. Sci. 2017, 1, x FOR PEER REVIEW 13 of 20
J. Compos. Sci. 2017, 1, x FOR PEER REVIEW 13 of 20

Figure 9. Walking test on the composite slab performed by Volunteer 3.

Figure 9. Walking
Walking test
test on
on the
the composite
composite slab
slab performed
performed by
by Volunteer
Volunteer3.3.

Figure 10. Response of Prototype I to walking performed by Volunteer 4 in terms of graphs: (a)
Figure 10.
Figure 10. Response
ResponseofofPrototype
PrototypeI Itotowalking
walkingperformed byby
performed Volunteer 4 in
Volunteer terms
4 in of graphs:
terms (a)
of graphs:
acceleration along time; (b) respective spectrum.
acceleration
(a) along
acceleration time;
along (b)(b)
time; respective spectrum.
respective spectrum.
Table 8. Accelerations from walking tests (m/s22).
Table8.8.Accelerations
Table Accelerationsfrom
fromwalking
walkingtests
tests(m/s ).
(m/s2 ).
High
High Peak
Peak Accelerations
Accelerations RMS
RMS Accelerations
Accelerations
Volunteer
Volunteer High Peak Accelerations RMS Accelerations
Volunteer Prototype I
Prototype I Prototype
Prototype II
II Prototype I
Prototype I Prototype
Prototype II
II
Prototype I Prototype II Prototype I Prototype II
1 11 5.27
5.27
5.27 4.86
4.86
4.86 0.86
0.86
0.86 0.83
0.83
0.83
22 4.37
4.37 3.49
3.49 0.95
0.95 0.52
0.52
2 4.37 3.49 0.95 0.52
3 6.43 6.10 1.25 1.09
43
3 6.43
6.43
5.03
6.10
6.10
4.85
1.25
1.25
1.28
1.09
1.09
1.30
54 5.03
5.56 4.85
3.51 1.28
1.41 1.30
0.94
4 5.03 4.85 1.28 1.30
Mean 5.33 4.56 1.15 0.94
55 5.56
5.56 3.51
3.51 1.41
1.41 0.94
0.94
Mean
Mean 5.33 4.56
5.33 high peak accelerations
4.56 1.15 0.94
Regarding human-sensitivity, between1.153.5 to 6.4 m/s2 (≈0.94 0.4 to 0.7 g) were
attained in both prototype responses, considering all the volunteers, which for 2a frequency around
Regarding human-sensitivity, high peak
peak accelerations between 3.53.5 to 6.4
6.4 m/s (
( 0.4
0.4 to 0.7
0.7Ing) were
12 HzRegarding human-sensitivity,
lie in the unbearable range ofhigh
vibration accelerations
perception inbetween
Goldman’s to
scale m/s
(see2 Figure to2). g)terms
were
attained
attained in
in both
both prototype
prototype responses,
responses, considering
considering all
all the
the volunteers,
volunteers,
2 which
which for
for aa frequency
frequency 2 around
around
of RMS accelerations, the results varied from 0.5 to 1.4 m/s , with an average of 1.05 m/s for both
12
12 Hz
Hz lie
lie in the
the unbearable
in These unbearable range of vibration perception in
in Goldman’s scale (see
(see Figure
Figure 2). In terms
prototypes. results arerange
aboveofthe
vibration perceptionlimit
RMS acceleration Goldman’s scale assessment)
(using general 2).
forIn terms
human
of
of RMS
RMS accelerations,
accelerations, the
the results
results varied
varied from
from 0.5
0.5 to
to 1.4
1.4 m/s
m/s
2, with an average of 1.05 m/s2 for both
2, with an average of 1.05 m/s2 for both
comfort by ISO 10137 [10], according to Figure 3.
prototypes.
prototypes. TheseThese results
results are
are above
above the
the RMS
RMS acceleration
acceleration limit
limit (using
(using general
general assessment)
assessment) for for human
human
comfort
4.2.3.
comfort by
by ISO
Jumping 10137
10137 [10],
ISOTest [10], according
according to to Figure
Figure 3.
3.
For this test,
4.2.3. the volunteers applied about 30 jumps in 10 seconds on the prototypes at the center,
4.2.3. Jumping
Jumping Test
Test
as shown in Figure 11, seeking to apply a jumping load with a frequency of around 3.0 Hz. Results from
For
For this
this test,
test, the
the volunteers
volunteers applied
applied about
about 30
30 jumps
jumps in
in 10
10 seconds
seconds on
on the
the prototypes
prototypes at
at the
the center,
center,
as shown in Figure 11, seeking to apply a jumping load with a frequency of around 3.0
as shown in Figure 11, seeking to apply a jumping load with a frequency of around 3.0 Hz. ResultsHz. Results
J. Compos. Sci. 2017, 1, 11 14 of 20
J. Compos. Sci. 2017, 1, x FOR PEER REVIEW 14 of 20
J. Compos. Sci. 2017, 1, x FOR PEER REVIEW 14 of 20
the test
from theperformed by Volunteer
test performed 4, in terms
by Volunteer 4, inofterms
response acceleration
of response in time and
acceleration frequency
in time domain,
and frequency
from
can bethe test
seen inperformed
Figure 12. by Volunteer
Table 9 4,
presents in
theterms
high of response
peak acceleration
accelerations and in
the
domain, can be seen in Figure 12. Table 9 presents the high peak accelerations and the RMS time
RMS and frequency
accelerations
domain, from
obtained can obtained
accelerations be
theseen in
testsfrom Figure
performed 12. Tablevolunteer.
byperformed
the tests each 9 presents the high peak accelerations and the RMS
by each volunteer.
accelerations obtained from the tests performed by each volunteer.

Figure 11. Jumping test on the composite slab performed by Volunteer 4.

Figure 11. Jumping
Figure 11. Jumping test
test on
on the
the composite
composite slab
slab performed
performed by
by Volunteer
Volunteer4.4.

Figure 12. Response of Prototype II to jumping performed by Volunteer 4 in terms of graphs: (a)
Figure 12.
Figure 12. Response
acceleration Response ofof(b)
along time; Prototype
Prototype IIIIto
respective tojumping
jumpingperformed
spectrum. byby
performed Volunteer 4 in
Volunteer terms
4 in of graphs:
terms (a)
of graphs:
acceleration
(a) along
acceleration time;
along time;(b)(b)
respective
respectivespectrum.
spectrum.
Higher peak accelerations (around 0.5 to 1.8 g) were attained as compared to the ones observed
in theHigher peak accelerations (around 0.5 to 12
1.8from
g) were
theseattained as compared to the ones observed
Table 9. Accelerations jumping tests (m/s 2 ).
walking tests. For a frequency of about Hz, accelerations are classified as unbearable
in the walking
according to thetests.
scaleFor a frequency
presented of about
in Figure 2. By12means
Hz, these accelerations
of RMS accelerations,are classified
the resultsasvaried
unbearable
from
according to the scale
1.2 to 5.0 m/sVolunteer presentedHighin Figure
Peak 2. By means
Accelerations of RMS accelerations,
RMS the
Accelerationsresults
2 for both prototypes. Again, all the values are above the limit recommended for human varied from
1.2 to 5.0by
m/s Prototype I Prototype II Prototype I Prototype II
ISOfor both[10],
prototypes. Again, all the
3. values are above the limit recommended for human
2
comfort 10137 as shown in Figure
comfort by ISO 110137 [10], as shown 9.82 in Figure 3.14.71 2.09 3.04
2 14.26
Table 9. Accelerations12.46
from jumping tests3.36
(m/s2). 2.69
3 Table
7.709. Accelerations11.42
from jumping tests (m/s2).
1.41 2.03
4 14.38Accelerations
High Peak 17.74 RMS 3.26
Accelerations 4.94
Volunteer
5 High Peak
5.09 Accelerations 10.83 RMS 1.19
Accelerations 2.06
Volunteer Prototype I Prototype II Prototype I Prototype II
Mean 10.25I
Prototype 13.43 II
Prototype 2.26 I
Prototype 2.95 II
Prototype
1 9.82 14.71 2.09 3.04
21 9.82
14.26 14.71
12.46 2.09
3.36 3.04
2.69
2
Higher peak
3 accelerations 14.26 12.46
(around 0.5 to 1.8
7.70 3.36
g) were attained1.41
11.42 as compared to the 2.69
2.03ones observed
in the walking43tests. For a frequency
7.70
14.38 of about 11.42
12 Hz,
17.74 these 1.41
accelerations
3.26 are 2.03
classified
4.94as unbearable
54
according to the scale 14.38
presented
5.09in Figure 2. By17.74
means
10.83 of RMS 3.26
accelerations,
1.19 the 4.94 varied from
results
2.06
5
Mean 5.09
10.25 10.83
13.43 1.19
2.26 2.06
2.95
Mean 10.25 13.43 2.26 2.95
J. Compos. Sci. 2017, 1, 11 15 of 20

J. Compos. Sci. 2017, 1, 11 15 of 19

1.2 to 5.0 Sci.

J. Compos. m/s 2 for
2017, 1, xboth
FOR prototypes.
PEER REVIEWAgain,
all the values are above the limit recommended for human 15 of 20
4.3. Numerical Analysis
comfort by ISO 10137 [10], as shown in Figure 3.
4.3. Numerical Analysis analysis of the slab prototypes, the Finite Element Method (FEM) was applied.
For the numerical
4.3. Numerical
The slab Analysis
prototypes were modeled
For the numerical analysis of theby
slabfinite elements
prototypes, theasFinite
shown in Figure
Element Method 13, (FEM)
with 8-node solid
was applied.
elements
The For (with
slabthe incompatible
numericalwere
prototypes Wilson
analysis modes)
of the slab
modeled used
prototypes,
by finite to
elements model
the as
Finitethe concrete
Element
shown top
Method
in Figure and
13,(FEM)4-node Discrete
was applied.
with 8-node solid
Kirchhoff
The slab Quadrilateral
prototypes were (DKQ)
modeled shell
by elements
finite elementsto model
as shownthe GFRP
in Figure profiles
13, with(the
elements (with incompatible Wilson modes) used to model the concrete top and 4-node Discrete foam
8-node blocks
solid were
elements
disregarded).
(with
Kirchhoff Care was
incompatible
Quadrilateral taken
Wilson in the
modes)
(DKQ) modeling
used
shell elements toto
to model represent accurately
the concrete
model the GFRP the same
top and 4-node
profiles boundary
(the Discrete conditions
Kirchhoff
foam blocks were
as those in
Quadrilateralthe experimental
(DKQ) shell analysis.
elements to The
modelanalyses
the were
GFRP carried
profiles out
(the using
foam SAP
blocks
disregarded). Care was taken in the modeling to represent accurately the same boundary conditions 2000
were [4] computer
disregarded).
program.
Care was in
as those taken
the in the modeling
experimental to represent
analysis. accurately
The analyses the same
were boundary
carried out using conditions
SAP 2000as[4] those in the
computer
experimental
program. analysis. The analyses were carried out using SAP 2000 [4] computer program.

Figure 13. Finite element model of the composite slab prototype.

Figure 13. Finite element model of the composite slab prototype.
13. Finite element model of the composite slab prototype.
Materials were Figure assumed to behave linear-elastically. The equivalent orthotropic elastic
properties of the GFRP profiles
Materials were assumed to behave utilized are shown in Table 1, and
linear-elastically. Thetheequivalent
elastic properties adopted
orthotropic for
elastic
Materials
concrete were were assumed
described to behave
in subsection linear-elastically. The equivalent orthotropic elastic properties
properties of the GFRP profiles utilized3.1.
areAshown
damping coefficient
in Table 1, and of
the4.96%,
elasticderived
propertiesfrom the heel-
adopted for
of
dropthetests,
GFRPwasprofiles utilized
adopted in are
all shown in Table 1, and the elastic properties adopted for concrete were
analyses.
concrete were described in subsection 3.1. A damping coefficient of 4.96%, derived from the heel-
described in subsectionmodels
3.1. A for
damping coefficient
loadsofused
4.96%, derived the
from the heel-drop tests, tests
was
dropThe mathematical
tests, was adopted in all analyses. the dynamic to simulate walking and jumping
adopted
were The in all
the ones analyses.
aforementioned
mathematical modelsdescribed. In bothloads
for the dynamic loadings,
used tothree harmonics
simulate (n = 3), and
the walking and jumping
the weight of
tests
each The mathematical
volunteer were models forThe
considered. the dynamic loads used toprocedure
mode-superposition simulate the
was walking
utilized and
in jumping
the tests
numerical
were the ones aforementioned described. In both loadings, three harmonics (n = 3), and the weight of
were the ones
analyses, aforementioned
considering described.
20 vibration modes. In both loadings, three harmonics (n = 3), and the weight of
each volunteer were considered. The mode-superposition procedure was utilized in the numerical
each The
volunteer were considered. The mode-superposition procedure was utilized in the numerical
analyses,natural frequencies
considering and respective
20 vibration modes. vibration modes of the slab prototype were obtained
analyses,
usingThe theconsidering
modal analysis 20 vibration
option modes.
natural frequencies andinrespective
SAP2000 [4] program.
vibration Figure
modes 14 displays
of the the firstwere
slab prototype threeobtained
natural
The
frequencies natural frequencies
foundanalysis and
for the option
compositerespective vibration
slab (11.63, modes of the slab prototype were obtained using
using the modal in SAP2000 [4] 26.00
program.and Figure
41.08 Hz) and theirthe
14 displays respective
first three vibration
natural
the modal
modes. analysis option in SAP2000 [4] program. Figure 14 displays the first three natural frequencies
frequencies found for the composite slab (11.63, 26.00 and 41.08 Hz) and their respective vibration
found for the composite slab (11.63, 26.00 and 41.08 Hz) and their respective vibration modes.
modes.
The response from all numerical analyses was extracted at the same position of the
accelerometers, i.e., at the slab midspan, in the center of the bottom flanges of the GFRP profiles.
In order to simulate the eight turns performed by the volunteers on the composite slab prototype
during the test, the walking load, described by Equations (2) to (9), was applied at eight positions
along the slab’s span, with a time interval of 0.5 s. Figure 15 shows the slab response for walking
obtained numerically, in terms of graphs of acceleration along time and its respective spectrum,
considering the weight of Volunteer 4. The acceleration (a) spectrum indicates that the structure
responds predominantly in the fundamental frequency, approximately 12 Hz. Peak accelerations of
3.6 m/s2 (around 0.4 g), and corresponding RMS acceleration of 0.97 m/s2 were obtained. These values
lie in the unbearable range of vibration perception in Goldman’s scale (Figure 2) for a structure with
fundamental frequency of 12 HZ. The RMS value is much higher than the limit given by ISO 10137
[10].
(b) (c)

Figure 14. Natural frequencies and respective vibration modes of the composite slab: (a) first mode (bending
Figure 14. Natural frequencies and respective vibration modes of the composite slab: (a) first mode
about y axis); (b) second mode (torsion about x axis) and; (c) third mode (bending about y axis).
(bending about y axis); (b) second mode (torsion about x axis) and; (c) third mode (bending about
y axis).

(a)
J. Compos. Sci. 2017, 1, 11 16 of 20

The response from all numerical analyses was extracted at the same position of the accelerometers,
J. Compos.
J. Compos. Sci.
Sci. 2017,
2017, 1,
1, xx FOR
FOR PEER
PEER REVIEW
REVIEW 16 of
16 of 20
20
i.e., at the slab midspan, in the center of the bottom flanges of the GFRP profiles.
In order to simulate the eight turns performed by the volunteers on the composite slab prototype
during the test, the walking load, described by Equations (2) to (9), was applied at eight positions along
the slab’s span, with a time interval of 0.5 s. Figure 15 shows the slab response for walking obtained
numerically, in terms of graphs of acceleration along time and its respective spectrum, considering the
weight of Volunteer 4. The acceleration spectrum indicates that the structure responds predominantly
in the fundamental frequency, 2
(b) approximately 12 Hz. Peak accelerations of 3.6(c)
(b) (c)m/s (around 0.4 g), and
2
corresponding RMS acceleration of 0.97 m/s were obtained. These values lie in the unbearable range
Figure
ofFigure 14. Natural
vibration
14. Natural
perceptionfrequencies and respective
in Goldman’s
frequencies and respective vibration
scale vibration
(Figure 2)modes of
for a of
modes the composite
composite
structure
the slab: (a)
(a) first
first mode
with fundamental
slab: mode (bending
frequency
(bending of
12 Hz. Theabout
about RMSy axis);
value
y axis); (b) second
(b) is mode
muchmode
second (torsion
higher about
than about
(torsion x axis)
the limit and;
given
x axis) and;by(c) third
(c)ISO mode
10137
third (bending
mode[10]. about y axis).
(bending about y axis).

Figure
Figure 15. Numerical response
15. Numerical
Numerical responseofofthe
thecomposite
compositeslab
slabtotowalking
walking loading
loading in in terms
terms of graphs:
of graphs:
graphs: (a)
Figure 15. response of the composite slab to walking loading in terms of (a)
(a) acceleration along
acceleration along time;
along time;
time; (b)(b) respective
(b) respective spectrum.
respective spectrum.
spectrum.
acceleration

The jumping
The
The jumping load
jumping load described
load described in
described in Section
in Section 222(obtained
Section (obtainedby
(obtained bymultiplying
by multiplying FF
multiplying Fiii(t),
(t), shown
(t), shown in
shown in Figure
in Figure 5,
Figure 5, by
5, by
by
the
the person’s
the person’s weight,
weight, G)
person’s weight, G) was
G) was applied
was applied distributed
applied distributed
distributedover over
overtwo two areas,
twoareas, corresponding
areas,corresponding
correspondingtotothe to the shoe/slab
theshoe/slab
shoe/slab
contact
contact area,
area, at
at the
the center
center of
of the
the numerical
numerical model,
model, in
in the
the
contact area, at the center of the numerical model, in the same position same
same position
position as
as applied
applied in
in the
the experiments.
experiments.
Theresults
The
The resultsfor
results for jumping
for jumping are
jumping are shown
are shown in
shown in Figure
in Figure 16,
Figure 16, for
16, for the
for the test
the test performed
test performed by
performed by Volunteer
by Volunteer4,4,
Volunteer 4,in
interms
in terms of
terms of
of
graphsof
graphs
graphs ofacceleration
of accelerationalong
acceleration alongtime
along timeand
time andits
and itsrespective
its respectivespectrum.
respective spectrum.From
spectrum. Fromthe
From theacceleration
the accelerationspectrum
acceleration spectrumitit
spectrum itcan
can
can
be
be observed,
observed, again,
again, that
that the
the slab
slab responds
responds primarily
primarily at
at its
its fundamental
fundamental mode.
mode.
be observed, again, that the slab responds primarily at its fundamental mode. The peak accelerations The
The peak
peak accelerations
accelerations
obtainednumerically
obtained
obtained numericallyreached
numerically reachedatat
reached atmost
most9.47
most 9.47m/s
9.47 m/s222with
m/s withRMS
with RMSof
RMS of3.29
of 3.29m/s
3.29 m/s222.. These
m/s These values
values are
are much
are much higher
much higher
higher
thanthe
than
than thelimit
the limitvalues
limit valuesgiven
values givenby
given byFigures
by Figures222and
Figures and3.3.
and 3.

Figure 16.
Figure 16. Numerical response
response of the composite
composite slab to
to jumping loading
loading in
in terms
terms of
of graphs:
graphs: (a)
(a)
Figure 16. Numerical
Numerical responseofofthe the compositeslab
slab tojumping
jumping loading in terms of graphs:
acceleration along
acceleration along time;
time; (b)
(b) respective
respective spectrum.
spectrum.
(a) acceleration along time; (b) respective spectrum.

5. Analysis
5. Analysis and
and Discussion
Discussion of
of Results
Results
The three
The three methods
methods used
used to
to estimate
estimate the
the fundamental
fundamental frequency
frequency of
of the
the composite
composite slab
slab yielded
yielded
similar values. The simplified theoretical method, based on beam analogy combined
similar values. The simplified theoretical method, based on beam analogy combined with the with the
Transformed Section
Transformed Section approach
approach resulted
resulted in
in aa value
value of
of 11.88
11.88 Hz;
Hz; the
the heel
heel impact
impact tests
tests yielded
yielded an
an
J. Compos. Sci. 2017, 1, 11 17 of 20

5. Analysis and Discussion of Results

The three methods used to estimate the fundamental frequency of the composite slab yielded
similar values. The simplified theoretical method, based on beam analogy combined with
J. Compos. Sci. 2017, 1, x FOR PEER REVIEW
the
17 of 20
Transformed Section approach resulted in a value of 11.88 Hz; the heel impact tests yielded an
average
averagevalue
valueof of 12.64
12.64 Hz,
Hz, while
while the
the numerical
numerical analysis
analysis based
based on on the
the Finite
Finite Element
Element Method
Method provided
provided
aa value of 11.63 Hz. Thus, it can be inferred that the fundamental
value of 11.63 Hz. Thus, it can be inferred that the fundamental frequency of the slab frequency of the slabprototype
prototype is
is about
about 1212Hz. Hz. According
According to EN
to EN 19901990
[9], [9], verification
verification of theof comfort
the comfort criteria
criteria shouldshould be performed
be performed if the
iffundamental
the fundamental frequency
frequency is lessis less
thanthan 5 Hz.
5 Hz. The The compositeslab
composite slabstudied
studiedhere here has
has aa fundamental
fundamental
frequency of 12 Hz, and therefore should not present any vibration
frequency of 12 Hz, and therefore should not present any vibration problem regarding problem regarding human comfort.
human
However, results of results
comfort. However, the experimental and numerical
of the experimental analyses have
and numerical shown
analyses theshown
have opposite.the opposite.
The
The experimental
experimentaland and numerical
numerical analyses
analyses indicated
indicated that that the
the composite
composite slab slab responded
responded primarily
primarily
in
in the
the first
first vibration
vibration modemode (flexural
(flexural mode),
mode), asas can
can bebe seen
seen from
from thethe frequency
frequency spectrum
spectrum shown shown in in
Figures 10, 12, 15 and 16. The natural frequencies for the second and third
Figures 10, 12, 15 and 16. The natural frequencies for the second and third modes of vibration (26.00 modes of vibration (26.00
and
and 41.08
41.08 Hz)
Hz) diddidnotnotshow
showsignificant
significantamplitudes.
amplitudes.
The
The comparison
comparison betweenbetween the the experimental
experimental and and the
the numerical
numerical results
results inin terms
terms of of accelerations
accelerations
obtained
obtained for the walking load from Volunteer 4 can be visualized in Figure 17 for the initial cycles.
for the walking load from Volunteer 4 can be visualized in Figure 17 for the initial cycles.
Despite
Despite thethe good
good agreement,
agreement, the the results
results from
fromthethe experimental
experimentaltest testin
in terms
terms ofof high
high peak
peak acceleration
acceleration
(5.03
(5.03 and
and 4.85 m/s2)2 )are
4.85 m/s arearound
around35% 35% greater
greater than
than thethe results
results from
from thethe numerical
numerical model
model (3.6(3.6
m/s2m/s
2 ).
). This
This difference is also observed for RMS
the RMS accelerations, when comparing 2
difference is also observed for the accelerations, when comparing 1.30 1.30 to 0.97
to 0.97 m/s2m/s, found, found
from
from the experimental
the experimental and numerical
and numerical analysis,
analysis, respectively.
respectively.

Figure 17.
Figure 17. Comparison
Comparison between
between acceleration
acceleration graphs
graphs obtained
obtained numerically
numerically and
and experimentally
experimentally from
from
the test on Prototype I performed by Volunteer 4 walking.
the test on Prototype I performed by Volunteer 4 walking.

For the jumping loading, the numerical results did not show a good approximation to the
For the jumping loading, the numerical results did not show a good approximation to the
experimental results for the initial cycles. Figure 18 displays the comparison between numerical and
experimental results for the initial cycles. Figure 18 displays the comparison between numerical and
experimental accelerations obtained from the test performed by Volunteer 4. Although the high peak
experimental accelerations obtained from the test performed by Volunteer 4. Although the high peak
accelerations from the numerical model are much smaller than the ones from the experimental test
accelerations from the numerical model are much smaller than the ones from the experimental test
(9.47 m/s2 as compared to 14.38 and 17.74 m/s2), the RMS accelerations are similar (3.29 m/s2 for the
(9.47 m/s2 as compared to 14.38 and 17.74 m/s2 ), the RMS accelerations are similar (3.29 m/s2 for the
numerical model and 3.26 and 4.94 m/s2 for the experimental test). This similarity in RMS values may
numerical model and 3.26 and 4.94 m/s2 for the experimental test). This similarity in RMS values may
be due to the difficulty of imposing a constant load rhythm on the slab by the volunteers, which does
be due to the difficulty of imposing a constant load rhythm on the slab by the volunteers, which does
not occur when the slab was analyzed numerically.
not occur when the slab was analyzed numerically.
J. Compos. Sci. 2017, 1, 11 18 of 20
J. Compos. Sci. 2017, 1, x FOR PEER REVIEW 18 of 20

Figure 18.
Figure 18. Comparison
Comparison between
between acceleration
acceleration graphs
graphs obtained
obtained numerically
numerically and
and experimentally
experimentally from
from
test on Prototype II performed by Volunteer 4 jumping.
test on Prototype II performed by Volunteer 4 jumping.

Nonetheless, from both the experimental and numerical analyses, the composite slab showed
Nonetheless, from both the experimental and numerical analyses, the composite slab showed
vibrations above the limit of human comfort according to Goodman’s scale (Figure 2) for peak
vibrations above the limit of human comfort according to Goodman’s scale (Figure 2) for peak
acceleration and also according to ISO 10137 [10] for RMS acceleration. This result indicates that the
acceleration and also according to ISO 10137 [10] for RMS acceleration. This result indicates that
comfort criteria should be taken into account in the slab design, even though its fundamental
the comfort criteria should be taken into account in the slab design, even though its fundamental
frequency is above 5 Hz.
frequency is above 5 Hz.
These results have shown that, in spite of their good performance under static loads, composite
These results have shown that, in spite of their good performance under static loads, composite
slabs made of concrete and GFRP profiles tend to be more flexible and slender, and therefore their
slabs made of concrete and GFRP profiles tend to be more flexible and slender, and therefore
dynamic behavior under human activities must be considered in the design. Regarding the particular
their dynamic behavior under human activities must be considered in the design. Regarding the
composed slab being developed, a greater efficiency is expected if wide flange GFRP profiles and/or
particular composed slab being developed, a greater efficiency is expected if wide flange GFRP profiles
reduced spans are utilized, since that would increase the slab rigidity and consequently its
and/or reduced spans are utilized, since that would increase the slab rigidity and consequently its
fundamental frequency.
fundamental frequency.
It should be noted that in small specimens, as the ones used in this work, the mass of the person
It should be noted that in small specimens, as the ones used in this work, the mass of the person
on the slab may have some influence in the results; also, for the walking case, it takes some time until
on the slab may have some influence in the results; also, for the walking case, it takes some time until
the structure “gets in rhythm” with the applying load.
the structure “gets in rhythm” with the applying load.
6. Conclusions
6. Conclusions
This work
This work investigated
investigated the the dynamic
dynamic behavior
behavior of of aa composite
composite concrete/GFRP
concrete/GFRP slab, slab, under
under
development at UFSC, subjected to human activities. By considering a representative
development at UFSC, subjected to human activities. By considering a representative slab strip, slab strip, two
prototypes
two werewere
prototypes constructed for the
constructed for experimental
the experimental program.
program.A numerical finitefinite
A numerical element analysis
element was
analysis
also performed. From the analyses, the following conclusions
was also performed. From the analyses, the following conclusions could be drawn: could be drawn:
The fundamental
The fundamentalfrequency
frequencyofofthethecomposite
compositeslab slabisisapproximately
approximately 1212 Hz.
Hz. Although
Although this
this value
value is
is more than twice the limit indicated by EN 1990 (5 Hz), below which a dynamic
more than twice the limit indicated by EN 1990 (5 Hz), below which a dynamic analysis is required, analysis is required,
it was
it was observed
observed thatthat the
the composite
composite slab
slab showed
showed vibrations
vibrations above
above the
thelimit
limitfor
forhuman
humancomfort.
comfort.
 The slab responds primarily at its first vibration mode (flexural). The other vibration modes did
• not show
The significant
slab responds amplitudes.
primarily at its This indicates mode
first vibration that in(flexural).
a simplifiedTheapproach, even considering
other vibration modes did
justshow
not the first mode inamplitudes.
significant the analysis,Thisshould lead tothat
indicates goodin results.
a simplified approach, even considering
 Results
just frommode
the first experimental tests in
in the analysis, termslead
should of to
accelerations
good results. were always higher than the ones
• obtainedfrom
Results fromexperimental
the numericaltestsmodel. In addition,
in terms in the experimental
of accelerations tests the
were always peakthan
higher accelerations
the ones
did not follow a uniform pattern as the ones observed in the numerical
obtained from the numerical model. In addition, in the experimental tests the peak accelerations analysis. This may be
duenot
did to the factathat
follow the volunteers
uniform pattern aswere not observed
the ones able to maintain a constantanalysis.
in the numerical load rhythm.
This may be due
 to the fact that the volunteers were not able to maintain a constant load rhythm. requirements,
The composite slab was designed under static loads to attend both ULS and SLS
and it was verified experimentally under 4-point bending tests. However, in spite of its apparent
stiffness, with a fundamental frequency above 5 Hz, the concrete/GFRP composite slab under
study was shown to be sensitive to human activities. Therefore, in addition to the usual
J. Compos. Sci. 2017, 1, 11 19 of 20

• The composite slab was designed under static loads to attend both ULS and SLS requirements,
and it was verified experimentally under 4-point bending tests. However, in spite of its apparent
stiffness, with a fundamental frequency above 5 Hz, the concrete/GFRP composite slab under
study was shown to be sensitive to human activities. Therefore, in addition to the usual verification
of ultimate and serviceability limit states, the dynamic behavior of the slab must also be considered
in the design.

Acknowledgments: The authors would like to acknowledge the companies CSE Composites, SIKA, Termotécnica,
NeoNatex and Polimix for providing the material necessary to the experimental program. Thanks are also due to
FAPESC (Fundação de Apoio à Pesquisa Científica e Tecnológica do Estado de Santa Catarina) for the financial
support and to CNPq (Conselho Nacional de Desenvolvimento Científico e Tecnológico) for the scholarship
granted to the first author.
Author Contributions: Paulo Junges and Henriette Lebre La Rovere conceived and designed the experiments;
Paulo Junges performed the experiments; Paulo Junges, Henriette Lebre La Rovere and Roberto Caldas de
Andrade Pinto analyzed the data and wrote the paper.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Neto, A.B.S.; La Rovere, H.L. Composite concrete/GFRP slabs for footbridge deck systems. Compos. Struct.
2010, 92, 2554–2564. [CrossRef]
2. Junges, P. Vibration analysis of a composite slab of concrete and GFRP profiles induced by human activities.
Master dissertation, Civil Engineering Department, Federal University of Santa Catarina, Florianópolis,
Brazil, 2010. (In Portuguese)
3. Pinto, R.C.A.; Vieira, D.; La Rovere, H.L. Composite Concrete/GFRP Slabs under Concentrated Loads. In
New Developments in Structural Engineering and Construction; Research Publishing Services: Singapore, 2013;
Volume 2, pp. 1117–1122.
4. SAP2000. Integrated Finite Elements Analysis and Design of Structures, 11th ed.; Computers and Structures, Inc.:
Berkeley, CA, USA, 2009.
5. Živanović, S.; Pavic, A.; Reynolds, P. Vibration serviceability of footbridges under human-induced excitation:
A literature review. J. Sound. Vib. 2005, 279, 1–74. [CrossRef]
6. Bachmann, H. Case Studies of Structures with Man-Induced Vibrations. ASCE J. Struct. Eng. 1992, 118,
631–647. [CrossRef]
7. Reiher, H.; Meister, F.J. Die Empfindlichkeit dês Menschen gegen Erschütterungen (in German) (The Effect
of Vibration on People). Forsch. Ing. Wes. 1931, 2, 381–386.
8. Goldman, D.E. USN MRI Rept. 1, NM 004 001. March 1948.
9. EN 1990. Eurocode 0: Basis of Structural Design; European Committee for Standardization (CEN): Brussels,
Belgium, 2002.
10. ISO 10137: Bases for Design of Structures-Serviceability of Buildings and Walkways Against Vibrations; International
Organization for Standardization: Geneva, Switzerland, 2007.
11. Bachmann, H.; Pretlove, A.J.; Rainer, H. Dynamic forces from rhythmical human body motions. In Vibration
Problems in Structures: Practical Guidelines; Birkhäuser: Basel, Switzerland, 1995; Appendix G.
12. Varela, W.D. Theoretical-experimental model to analyze vibrations induced by people walking on floor slabs
of buildings. D.Sc. Thesis, Federal University of Rio de Janeiro, Rio de Janeiro, Brazil, 2004. (In Portuguese)
13. Figueiredo, F.P.; Silva, J.G.S.; Lima, L.R.O.; Vellasco, P.C.G.S.; Andrade, S.A.L. A parametric study of
composite footbridges under pedestrian walking loads. Eng. Struct. 2007, 30, 615–630. [CrossRef]
14. Mello, A.V.A.; Silva, J.G.S.; Vellasco, P.C.G.S.; Andrade, S.A.L.; Lima, L.R.O. Dynamic analysis of composite
systems made of concrete slabs and steel beams. J. Constr. Steel Res. 2008, 64, 1142–1151. [CrossRef]
15. Rainer, J.H.; Pernica, G.; Allen, D.E. Dynamic Loading and Response of Footbridges; Institute for Reseach in
Construction, National Research Council of Canada: Ottawa, ON, Canada, 1988.
16. Ohlsson, S.V. Floor vibrations and human disconfort. Ph.D. thesis, Departament of Structural Engineering,
Chalmers University of Technology, Götemborg, Sweden, 1982.
J. Compos. Sci. 2017, 1, 11 20 of 20

17. David, D.; Tristão, G.; Becocci, L.; Juliani, M.A. Experimental and computational analysis of floors submitted
to dynamic loads. In Proceedings of the 3rd International Operational Modal Analysis Conference-IOMAC09,
Portonovo, Italy, 4–6 May 2009; pp. 391–400.
18. Santos Neto, A.B.S. Development of a composite concrete/GFRP slab for footbridge deck systems.
Ph.D. Thesis, Civil Engineering Dept, Federal University of Santa Catarina, Florianópolis, Brazil, 2006.
(In Portuguese)
19. NBR 14859-1 (2002): Laje pré-fabricada-Requisitos. Parte 1: Lajes unidirecionais; Brasileira de Normas Técnicas
(ABNT): Rio de Janeiro, Brazil, 2002. (In Portuguese)
20. Deskovic, N.; Triantafillou, T.C.; Meier, U. Innovative design of FRP combined with concrete: Short-term
behavior. J. Struct. Eng. 1995, 121, 1069–1078.
21. Kaw, A.K. Mechanics of Composite Materials, 1st ed.; CRC Press: New York, NY, USA, 1997.
22. Bank, L.C. Flexural and shear moduli of full-section fiber reinforced plastic (FRP) pultruded beams. J. Test.
Eval. 1989, 17, 40–45.
23. La Rovere, H.L.; Santos Neto, A.B.S. Flexural stiffness characterization of fiber reinforced plastic (FRP)
pultruded beams. Compos. Struct. 2007, 81, 274–282.
24. ACI 318-11 (2011): Building Code Requirements for Structural Concrete; American Concrete Institute (ACI):
Farmington Hills, MI, USA, 2011.
25. Clougn, R.W.; Penzien, J. Dynamics of Structures, 2nd ed.; McGraw-Hill Education: New York, NY, USA, 1993.
26. CAN3-S16.1-M84: Steel Structures for Buildings-Limits States Design, Appendix G: Guide on Floor Vibrations;
Canadian Standards Association: Mississauga, ON, Canada, 1984.

© 2017 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/).