applied

sciences
Article
Bridge Expansion Joint in Road Transition Curve:
Effects Assessment on Heavy Vehicles
Paola Di Mascio, Giuseppe Loprencipe *, Laura Moretti, Lorenzo Puzzo and Pablo Zoccali
Department of Civil, Constructional and Environmental Engineering, Sapienza University of Rome,
Via Eudossiana 18, 00184 Rome, Italy; paola.dimascio@uniroma1.it (P.D.M.); laura.moretti@uniroma1.it (L.M.);
lorenzo.puzzo@uniroma1.it (L.P.); pablo.zoccali@uniroma1.it (P.Z.)
* Correspondence: giuseppe.loprencipe@uniroma1.it; Tel.: +39-064-458-5112

Academic Editor: Zhanping You

Received: 11 May 2017; Accepted: 7 June 2017; Published: 9 June 2017

Abstract: Properly-designed road surfaces provide a durable surface on which traffic can pass
smoothly and safely. In fact, the main causes that determine the structural decay of the pavement and
its parts are the traffic loads. These repeated actions can create undesirable unevennesses on the road
surface, which induce vertical accelerations on vehicles, up to hindering contact between pavement
and tire, with dangerous consequences on traffic safety. The dynamic actions transmitted by the
vehicles depend on these irregularities: often, a bridge expansion joint (BEJ), introducing a necessary
discontinuity between different materials, determines from the beginning a geometric irregularity in
the running surface. Besides, some structural conditions could emphasize the problem (e.g., local
cracking due to the settlement of the subgrade near the abutment or the discontinuity of stiffness due
to the presence of different materials). When the BEJ is located in a transition curve, an inevitable
vertical irregularity between road and joint can reach values of some centimeters, with serious
consequences for the road safety. This paper deals with the analysis of a case study of a BEJ. Several
test surveys were performed in order to fully characterize the effects on both vehicles and pavement.
The three-dimensional representation of the pavement surface and the acceleration measurements
on a heavy test vehicle were performed to analyze the joint behavior under traffic. Finally, a finite
element model was implemented to evaluate the stress contribution on vehicle components induced
by the vertical irregularities.

Keywords: bridge joint; heavy vehicles; road unevenness; road safety

1. Introduction
Highway bridge expansion joints (BEJ) allow vehicles to move between two uncoupled, but even,
adjacent road sections. Generally, a road built on a structure needs a longitudinal break to reduce
the effects of hygrothermal and/or seismic deformations and tangential stresses due to the traffic.
Sometimes, complex structure configurations of a bridge (e.g., arch bridge) permit avoiding BEJs,
in addition assuring aesthetic results for the artwork [1]. BEJ should absorb possible displacements of
the two parts, ensure bending and shear transmission and guarantee safe traffic circulation in every
condition (dry, wet, warm, freeze, etc.) during the service life [2]. Moreover, they should absorb
all actions induced by the relative displacements of coupled structures: thermal variations, wheel
dynamic vertical loads, braking and other horizontal forces [3].
BEJ presents a planarity problem when the joint is located in a transaction section of the
road (transition curve-clothoid) between two consequent road elements with different curvatures.
In fact, in this section, the road transversal slope is changing between the values necessary in
two adjacent elements with different horizontal curvatures. Consequently, in this transitional area,
the road pavement is characterized by a skew or slanted surface (Figure 1) and, if the joint is large

Appl. Sci. 2017, 7, 599; doi:10.3390/app7060599 www.mdpi.com/journal/applsci

Appl. Sci. 2017, 7, 599 2 of 21
Appl. Sci. 2017, 7, 599 2 of 20

is characterized by a skew or slanted surface (Figure 1) and, if the joint is large (e.g., a highway seismic
(e.g., a highway seismic joint), the vertical irregularity between road and joint can reach values of
joint), the vertical irregularity between road and joint can reach values of some centimeters.
some centimeters.
This undesirable local irregularity in addition to the construction unevenness can cause
This undesirable local irregularity in addition to the construction unevenness can cause important
important dynamic actions on running vehicles, the bridge structure and joints [4,5].
dynamic actions on running vehicles, the bridge structure and joints [4,5].

CROSS
SLOPE 2

CROSS
SLOPE 1
BEJ

IRREGULARITY

BEJ DIMENSION

Figure 1.
Figure Irregularity between
1. Irregularity between bridge
bridge expansion
expansion joint
joint (BEJ)
(BEJ) and
and the
the road
road in
in aa transition
transition curve.
curve.

In this case, dynamic actions induced by vehicle tires stress BEJ. Traffic running on bridges
produces a stress stress spectrum,
spectrum, which
which maymay cause
cause fatigue.
fatigue. The stress spectrum depends on various
parameters: the geometry of the vehicles,
parameters: the geometry of the vehicles, the axle loads, the axletheloads,
vehicle the vehicle
spacing, spacing,
superficial superficial
characteristics
characteristics
of the pavement of (unevenness)
the pavement[6], (unevenness)
the composition [6], the of composition
the traffic and of its
thedynamic
traffic and its dynamic
effects. Several
effects. Several
repetitions repetitions
of these dynamicofloads
theseduring
dynamic the loads
designduring design lifecan
life of the pavement of the pavement
cause a rapid andcan general
cause a
rapid and
decay of the general decay
structure of thebreaks
or local structure or local breakswith
in correspondence in correspondence
the BEJ, although with the BEJ, although
well-constructed [7].
well-constructed
A BEJ is anyhow [7]. A BEJ is anyhow
a discontinuity of thea discontinuity
pavement, and of dynamic
the pavement,
actionsand dynamic
caused actionsvehicles
by heavy caused
by heavyon
running vehicles running
it cannot on it cannotFor
be overlooked. be this
overlooked.
reason, in ForEurocode
this reason,
1 [8],in where
Eurocode 1 [8],
traffic where
loads on
traffic loads
bridges on bridges
are described, it isare described,
specified that init the
is specified
proximitythat in the proximity
of expansion joints, anof expansion
additional joints,
dynamic
an additional factor
amplification dynamic amplification
should be applied. factor should
In this sense,bethe applied. In this sense,
same Eurocode also the same aEurocode
suggests also
conservative
simplification consisting of adopting an amplification factor of 1.3 for any cross section within 6any
suggests a conservative simplification consisting of adopting an amplification factor of 1.3 for m
cross the
from section withinjoint.
expansion 6 m from the expansion joint.
Following specific recommendations for the fatigue design of modular bridge expansion joints [7],
it is possible
possible to to avoid
avoid BEJs’
BEJs’durability
durabilityproblems.
problems.InInany anycase,
case,the
theunevenness
unevenness induced
induced byby thethe
presence
presence of
a BEJ
of stillstill
a BEJ remains,
remains,andand
in certain conditions
in certain conditions(i.e.,(i.e.,
BEJ BEJ
located in road
located transition
in road curve),
transition it canitworsen.
curve), can worsen.
In particular, this paper focuses on the analysis of the detrimental effects induced by BEJ
unevenness in terms of road safety, safety, dynamic
dynamic increment
increment of wheel wheel loads transmitted to the pavement
and vehicle components’ stress.

2. Heavy Vehicle-Bridge
Vehicle-Bridge Interaction:
Interaction: The Expansion
Expansion Joint
Joint Role
Role
Moving heavy vehicles apply to roads and bridges wheel (dynamic) loads, which are higher
than nominal (static) loads. Static loads depend on the total weight of the vehicle and its geometric
axle configuration, whereas dynamic wheel forces depend on the road pavement profile, functional
characteristics ofofthe
thevehicle
vehicle(e.g., geometry,
(e.g., geometry,mass
massandand
stiffness distribution,
stiffness tire and
distribution, tire suspension type,
and suspension
operative speed) and structural characteristics of the bridge superstructure (i.e., span length,
type, operative speed) and structural characteristics of the bridge superstructure (i.e., span length, geometry,
static scheme,
geometry, natural
static scheme,frequencies and damping).
natural frequencies and damping).
Unevennesses due to a road BEJ can produce important dynamic load increments: this depends
on the
theinteraction
interactionbetween
betweenrunning vehicles
running and the
vehicles andbridge
the superstructure, as well asas
bridge superstructure, thewell
interactions
as the
between structural
interactions between components and vehiclesand
structural components [9].vehicles
Figure 2[9].
shows the2dynamic
Figure shows the interaction (coupling)
dynamic interaction
between
(coupling) vehicle
betweenandvehicle
bridge and
[10].bridge [10].
Appl. Sci. 2017, 7, 599 3 of 21
Appl. Sci. 2017, 7, 599 3 of 20

Initial Surface Dynamic Surface Roughness Vehicle

+
Roughness + (Input to the vehicle) Dynamics

Dynamic Tyre
Force
Dynamic Bridge Bridge
Deflection Dynamics

Figure 2. Schematic
Figure block
2. Schematic diagram
block diagramofofdynamic bridge-vehicleinteraction.
dynamic bridge-vehicle interaction.

The roughness input to the vehicle is the sum of the initial surface profile of the pavement and
The roughness input to the vehicle is the sum of the initial surface profile of the pavement and
the dynamic deflection of the bridge. This input excites the vehicle and results in dynamic tire forces.
the dynamic deflection of the bridge. This input excites the vehicle and results in dynamic tire forces.
Finally, these forces are in turn applied to the bridge and cause larger dynamic displacements of
Finally, these forces are in turn applied to the bridge and cause larger dynamic displacements of the
the bridge [11].
bridge [11]. The largest bridge vibration occurs when the bridge’s and vehicle’s natural frequencies are
The largest
similar [12]. bridge
It is notvibration occursif when
clear, however, the bridge’s
large bridge responses andarevehicle’s
mainly due natural frequencies
to excitation of the are
similarvehicle
[12]. Itcaused
is notbyclear, however,
the initial if largeprofile,
roughness bridgeorresponses are mainly
by the dynamic due of
deflection tothe
excitation of thewas
bridge. There vehicle
causedlittle
by systematic
the initial study concerning
roughness profile,the or
conditions when dynamic
by the dynamic interaction
deflection of theisbridge.
important and was
There whenlittle
the two
systematic systems
study may be regarded
concerning as essentially
the conditions when uncoupled.
dynamic interaction is important and when the two
systems may be regarded as essentially uncoupled.a fundamental natural frequency f ≅ 2.5–4 Hz
According to Cantieni [13], bridges with
(maximum span L ≅ 25–40 m) are more susceptible to the dynamic actions of heavy commercial
According to Cantieni [13], bridges with a fundamental natural frequency f ∼ = 2.5–4 Hz (maximum
vehicles
∼ (equipped with steel leaf suspensions) than bridges with a fundamental frequency lying
span L = 25–40 m) are more susceptible to the dynamic actions of heavy commercial vehicles (equipped
outside this range. This difference is because the wheel loads of such vehicles have a predominant
with steel leaf suspensions)
frequency than range
content in the same bridges [13].with
This aeffect
fundamental
is known asfrequency lying outside
“frequency-matching”: this range.
it leads to
This difference is because the wheel loads of such vehicles have a predominant
dynamic amplification of the corresponding strong dynamic bridge response, a phenomenon frequency content
also in
the same
knownrange [13]. This effect is known as “frequency-matching”: it leads to dynamic amplification of
as “quasi-resonance”.
the corresponding
The DIVINE strong dynamic
(Dynamic bridge response,
Interaction a phenomenon
between Vehicles also known
and Infrastructure as “quasi-resonance”.
Experiment) research
[14]DIVINE
The correlated the dynamic
(Dynamic bridge response
Interaction between under trafficand
Vehicles to the length-span bridge:
Infrastructure Experiment) research [14]
correlated the dynamic
1. Long-span bridge
element response
of bridges (L >under
100 m):traffic to the
The type length-span
of vehicle bridge:
suspension is not very significant
unless the road profile is poor. Furthermore, the fundamental frequency of bridges is less than
1. Long-span
that of element of bridges
heavy vehicles, and (L > 100
thus, m):bridges
these The type tendoftovehicle suspension
attenuate the dynamic is not very significant
components of
unless the road
wheel forces. profile is poor. Furthermore, the fundamental frequency of bridges is less than
that
2. of heavy vehicles,
Medium-span bridges ≅ 30–100
and(Lthus, these m):
bridges tend to attenuate
Frequency-matching with thethe
dynamic components
truck body bounce of
wheelfrequencies
forces. occurs in medium-span bridges. The experimental dynamic responses were well
2. within codebridges
Medium-span (L ∼
recommendations= 30–100 for high-quality road profiles.
m): Frequency-matching with the truck body bounce
3. Short to medium-span bridges (L ≅ 15–30 m): Bridges
frequencies occurs in medium-span bridges. The experimental dynamic with natural frequencies within thewere
responses rangewell
4–8 Hz tend not to dynamically couple with vehicles at either body bounce or axle-hop frequencies,
within code recommendations for high-quality road profiles.
respectively characterized by frequency within the range of 1.5–4.0 and 8.0–15.0 Hz [15].
3. Short to medium-span ∼
4. Short-span bridges (L bridges
≅ 8–15 m): (L “Quasi-resonance”,
= 15–30 m): Bridges with
but with natural
minimal frequencies
amplification, within
occurs
the range
between short-span bridges and the axle-hop vibrations of heavy vehicles. The effects of body or
4–8 Hz tend not to dynamically couple with vehicles at either body bounce
axle-hop
bouncefrequencies, respectively
are also transmitted characterized
to short-span bridgesby with
frequency within
minimal the range
amplification. Theof dynamic
1.5–4.0 and
8.0–15.0 Hz [15].
response of short span bridges can be much larger than expected for dynamic effects.
4. Short-span bridges
The DIVINE (L ∼
= 8–15
research has m):
shown “Quasi-resonance”,
that road profile, but with
bridge minimal
natural amplification,
frequency and bridge occurs
between
damping short-span
are important bridges andtothe
elements be axle-hop
controlled.vibrations
Road profile of measurement
heavy vehicles. Theroutine,
is now effects ofandbody
bounce are also transmitted
signal processing technologies areto relatively
short-span bridges withThe
straightforward. minimal
DIVINEamplification. The dynamic
project also demonstrated
that dynamic
response vehicle
of short spanmodels can accurately
bridges can be much reproduce
larger dynamic wheel forces,
than expected and many
for dynamic researchers
effects.
have demonstrated that the bridge/vehicle interaction problem can be analytically solved [12]. Some
The DIVINE
tools permit research
developinghasand
shown that road
validating profile,
bridge bridge
models naturalboth
including frequency and bridge
road profile damping
and bridge
natural frequency
are important [16].
elements to be controlled. Road profile measurement is now routine, and signal
processing technologies are relatively straightforward. The DIVINE project also demonstrated that
dynamic vehicle models can accurately reproduce dynamic wheel forces, and many researchers have
demonstrated that the bridge/vehicle interaction problem can be analytically solved [12]. Some tools
permit developing and validating bridge models including both road profile and bridge natural
frequency [16].
Appl. Sci. 2017, 7, 599 4 of 21

The model proposed by Green and Cebon [10] consists of a single degree of freedom vehicle
on a simply-supported bridge. It does not detail the nonlinearities of vehicle suspensions and the
complexities of sprung mass motion, but it ensures a good representation of the problem.
Cantieni [13] and Tan et al. [17] elaborated multiple-degree-of-freedom models to describe
dynamic bridge-vehicle interaction and found good agreement with experiments carried out on
highway bridges. In these approaches, road unevenness is modeled by artificial profiles that simulate
the desired roughness level [18].
Green and Cebon [10] performed a parametric study based on previously literature works [19–21]
and using the model proposed by Cantieni [13]. The results of this study permit ascertaining when
coupling is important and when it can be ignored. For highway bridges, the most important parameters
are speed, frequency and initial vehicle excitation. Moreover, vehicle suspensions are important in
determining the dynamic response of bridges under a single vehicle, as demonstrated both by the
DIVINE results and by Woodrooffe et al. [22]. Different suspension systems induce different dynamic
loads on roads (i.e., amplitude and frequency content). For example, for roads in fair condition and at
a speed of 80 km/h, the dynamic load induced by a truck fitted with airbag suspension is about 60%
of the dynamic load induced by the same truck, but fitted with steel multi-leaf suspension [22].
This paper presents a case study of a highway BEJ with a big unevenness due to the torsion of
the cross-section along a transition curve. After five years of service life, it had partial breaks and
superficial cracks in the head cross beam where it was located. Asphalt pavement before and after
the BEJ was affected by severe distresses (e.g., cracking and rutting). Indeed, the impulsive vibrating
actions of the heaviest and quickest vehicles produced a surface wear of the joint so important that
its service condition was very unsafe. The geometric configuration of the road section is such that
the allowable vehicle speed is quite high, and the concomitant presence of the joint determined the
cause of many accidents, even serious ones, which have affected both light and heavy vehicles [23].
The irregular circulation induced by the BEJ was causing hard braking and acceleration, which have
environmental effects [24]. Moreover, the condition of the BEJ had required several maintenance works
during its even short service life.
The present work focuses on the geometric characterization of the BEJ and on the assessment of
the effects on road traffic, with particular attention to heavy vehicles running on it. Specifically, safety
and comfort analyses were carried out, evaluating also the dynamic increment of loads transmitted to
the pavement (and then to the bridge) and the stress magnitude on vehicles’ mechanical components.
The importance of these analyses is that the interpretation of their results is useful to identify strategies
and procedures for policies to guarantee minimum safety levels at roads.

3. Methodology and Case of Study

The study of the bridge-vehicle interaction requires measuring and analytical treatment of the
road profile and the dynamic wheel forces applied to the investigated BEJ.
The profile analysis allows evaluating general ride characteristics referring to standards, comfort
and safety [25] and choosing pavement sections where the warning values of the previous evaluation
are reached and maintenance works needed [26–29].
The evenness properties of the road pavement profile can be measured using three different
methodologies:

• Topographic equipment: the rod and level are surveying tools used to represent the
three-dimensional pavement surface.
• Inclinometer-based profiling systems: they create a reference plan to measure the relative slope
during the equipment movement (Dipstick developed by The Face Companies, USA, “CHLOE”
profilometer, a device developed at the road test for determination of the slope variance on
pavements, USA) [30].
Appl. Sci. 2017, 7, 599 5 of 21

• Inertial profilometer: this equipment uses an internal inertial reference system (e.g., Longitudinal
Profile Analyzer “APL” developed by French Road Research Laboratory, Automatic Road
Analyzer “ARAN” developed by Fugro Roadware, Mississauga, Canada, etc.) [30].

The acceleration
Appl. Sci. 2017, 7, 599 analysis permits measuring dynamic tire forces and ride quality at5 different
of 20
speeds. Moreover, it permits determining the main vehicle oscillation and vibration components in the
The acceleration analysis permits measuring dynamic tire forces and ride quality at different
frequency domain and comparing them to the first bridge’s natural frequency. This kind of analysis
speeds. Moreover, it permits determining the main vehicle oscillation and vibration components in
can be performed by means of in situ measurements [31] and simulation models [32]: both approaches
the frequency domain and comparing them to the first bridge’s natural frequency. This kind of
were analysis
used in can the be
present study.
performed by means of in situ measurements [31] and simulation models [32]: both
approaches were used in the present study.
3.1. Case Study: An Unevenness Bridge Joint
3.1. Case
The caseStudy:
studyAn Unevenness
refers Bridge
to a steel Joint bridge characterized by a 50 m-long span (medium-span
girder
bridge). According to experimental
The case study refers to a steelresults published
girder bridge by Kim by
characterized et aal.50[12] for aspan
m-long span equal to 40.4 m,
(medium-span
bridge). According to experimental results published by Kim et al. [12] for
for this type of bridge, the fundamental natural frequencies of the first bending and torsion a span equal to 40.4 m,
modes
for this type of bridge, the fundamental natural frequencies of the first bending
(respectively 2.33 and 3.86 Hz) are within the characteristic frequency range of heavy vehicles’ and torsion modesbody
bounce(respectively 2.33 and
(1.5–4.0 Hz). For 3.86
this Hz) are within
reason, the characteristic
it is important to avoidfrequency range ofofheavy
the presence vehicles’
localized body
irregularities
bounce (1.5–4.0 Hz). For this reason, it is important to avoid the presence of localized irregularities
that could increase the loads transmitted to the bridge structure during the passages of heavy vehicles
that could increase the loads transmitted to the bridge structure during the passages of heavy
on the bridge structure. Furthermore, the examined highway carries a high percentage of daily
vehicles on the bridge structure. Furthermore, the examined highway carries a high percentage of
heavy vehicle
daily heavytraffic (Table
vehicle 1),(Table
traffic whose 1),effective traveling
whose effective speed speed
traveling was found to beto
was found within the range
be within the of
80–120 km/h.
range of 80–120 km/h.

Table1.1.Examined
Table Examined road trafficfeatures.
road traffic features.

Type of SpeedSpeed
Type of Road RangeRange
(km/h) Average
Average Daily
Daily Traffic
Traffic (ADT) % of
% Heavy
of Heavy Vehicles
Road (km/h) (ADT) Vehicles
Highway 80–120 20,500 20
Highway 80–120 20500 20

Inlast
In the the last
fivefive years,
years, anan accidentrate
accident ratehigher
higher than
than the
thenational
nationalaverage value
average [33][33]
value waswas
found for for
found
the examined Italian highway. Many accidents were localized in correspondence
the examined Italian highway. Many accidents were localized in correspondence to a bridge to a bridge and, in and,
particular, near one of the BEJs present in this area. After a brief in situ inspection, it was found that
in particular, near one of the BEJs present in this area. After a brief in situ inspection, it was found that
a reduction of the contact area between vehicles’ tires and pavement took place in many cases
a reduction of the contact area between vehicles’ tires and pavement took place in many cases during
during the passage on this BEJ, and sometimes, there was a complete loss of contact, as shown in
the passage
Figure 3.on this BEJ, and sometimes, there was a complete loss of contact, as shown in Figure 3.
As aAsconsequence
a consequenceof these issues,
of these issues,additional
additional in situmeasurements
in situ measurements were
were carried
carried out out to evaluate
to evaluate
the detrimental effects
the detrimental on on
effects thethe
bridge
bridgestructure,
structure, pavement, passengers
pavement, passengers and
and vehicles
vehicles induced
induced by theby the
peculiar geometry
peculiar of the
geometry BEJ.
of the BEJ.

Figure
Figure 3. Loss
3. Loss of of tire/pavement contact
tire/pavement contactduring
duringthe passage
the on on
passage the the
BEJ.BEJ.
Appl. Sci. 2017, 7, 599 6 of 21
Appl. Sci. 2017, 7, 599 6 of 20

Furthermore, the surface pavement surrounding the BEJ presented several distresses, caused by
the dynamic load increment of road
road vehicles
vehicles (Figure
(Figure 4).
4).

Figure
Figure 4.
4. Pavement
Pavement condition
condition of
of the
the area
area surrounding the BEJ.
surrounding the BEJ.

3.2. In Situ Test and Equipment

3.2. In Situ Test and Equipment
The following tests were performed on the highway section including the joint:
The following tests were performed on the highway section including the joint:
1. Precision geometric-topographic measurements, along with about 500 m-long branch: The survey
1. was Precision
madegeometric-topographic
using a precision theodolitemeasurements, along with
(total station) andabout 500assuming
a rod, m-long branch: The survey
the carriageway
was made using a precision theodolite (total station) and a rod, assuming
markings as the edges of the surveying area. The coordinate reference system was relative; just the carriageway
markings
the as the
elevations edges
were of the
linked surveying
to the absolute area. The coordinate
system. In the area reference
closest to system was
the joint, a 1relative; just
× 1 m mesh
the elevations were linked to the absolute system. In the area closest to
size was considered, while in the other distant zones, 3 m × 3 m and 10 m × 10 m meshes were the joint, a 1 × 1 m mesh
size was considered,
created. while in the
This discretization other distant
allowed performingzones,the × 3 m layout
3 mglobal and 10 m of×the
10 m meshes were
platform and,
created. Thisthe
particularly, discretization
carriageway allowed
markingsperforming
and the theroad global
edges.layout of the platform and, particularly,
the carriageway
2. Evaluation markings
of dynamic anddue
action the to
road
theedges.
joint, with accelerometers installed on a heavy vehicle
2. test:
Evaluation of dynamic action due to
The measured accelerations were analyzed the joint, with
in theaccelerometers installed
time and frequency on a heavy
domains. vehicle
In particular,
test: The measured accelerations
the equipment used was composed of: were analyzed in the time and frequency domains. In particular,
the equipment used was composed of:
• Three-axle heavy vehicle, with an overall empty load equal to 16.04 t (P1 = 3.208 t;
• P2 = P3 = 6.416
Three-axle t; seevehicle,
heavy Figure 5);
with an overall empty load equal to 16.04 t (P1 = 3.208 t;
• Three
P2 = P3concrete
= 6.416blocks whose5);
t; see Figure weights were, respectively, 1.82 (C1), 2.04 (C2) and 2.18 t (C3);
•• SevenThree concrete blocks whose weights4 were,
piezoelectric accelerometers: accelerometers
respectively,with
1.82a (C1),
bottom 2.04scale
(C2)±50
andg2.18
(Type A)
t (C3);
• and 3 with a minimum full-scale range of ±20 g (Type B). The Type B
Seven piezoelectric accelerometers: 4 accelerometers with a bottom scale ±50 g (Type A) andaccelerometers were
located
3 with aon the rightfull-scale
minimum end of the axles
range of connected
±20 g (Type toB).
theThe
vehicle
Type frame, while three
B accelerometers of the
were four
located
Type A accelerometers were placed one on each vehicle axle. The last Type
on the right end of the axles connected to the vehicle frame, while three of the four Type A A accelerometer
was located on were
accelerometers the left sideone
placed of the frontvehicle
on each axle toaxle.
control
Thethelastpossible asymmetry ofwas
Type A accelerometer the
transversal acceleration
located on the left side ofdistribution.
the front axle to control the possible asymmetry of the transversal
acceleration distribution.
 Appl. Sci. 2017, 7, 599 7 of 21
Appl. Sci. 2017, 7, 599 7 of 20

Figure
Figure5.5.Geometric
Geometric(in
(incm)
cm)and
andload
loadcharacteristics
characteristicsofofthe
thetest
testvehicle.
vehicle.

Inside the vehicle cab, a signal control exchange and a PC were located to record and verify
Inside the vehicle cab, a signal control exchange and a PC were located to record and verify
collected data and to re-calibrate the equipment if necessary.
collected data and to re-calibrate the equipment if necessary.
Four different loading conditions were considered to examine the phenomenon. They represent
Four different loading conditions were considered to examine the phenomenon. They represent
various heavy vehicle configurations (Table 2) and various dynamic increments, because the
various heavy vehicle configurations (Table 2) and various dynamic increments, because the
phenomenon is sensitive to both the speed and load condition of heavy vehicles, as reported by
phenomenon is sensitive to both the speed and load condition of heavy vehicles, as reported by
Lu et al. [34]. The considered speed and load combinations gave to the BEJ energy values comparable
Lu et al. [34]. The considered speed and load combinations gave to the BEJ energy values comparable
to those reached on the structure under real service conditions.
to those reached on the structure under real service conditions.
Table 2. Considered load-speed combinations.
Table 2. Considered load-speed combinations.
Test Number Speed (km/h) Speed (m/s) Load Blocks Total Load (t)
Test Number Speed (km/h) Speed (m/s) Load Blocks Total Load (t)
1 107 29.72 C1, C2 and C3 22.08
21 107
84 29.72
23.33 C1, C2 and
C1, C2 and C3 C3 22.08
22.08
32 84
103 23.33
28.61 C1, C2 and C3
C3 22.08
18.22
3 103 28.61 C3 18.22
44 85
85 23.61
23.61 C3C3 18.22
18.22

3.3. 3D Finite Element Model

3.3. 3D Finite Element Model
In addition to the in situ measurements, a simulation model was implemented. In particular,
In addition to the in situ measurements, a simulation model was implemented. In particular, a 3D
a 3D model of the asphalt concrete pavement and the seismic BEJ was developed using LS-DYNA.
model of the asphalt concrete pavement and the seismic BEJ was developed using LS-DYNA. It allowed
It allowed evaluating the stress magnitude on the test heavy vehicle mechanical components (Figure 6)
evaluating the stress magnitude on the test heavy vehicle mechanical components (Figure 6) due to
due to the vertical irregularity that characterizes the BEJ. For this reason, the vehicle is modeled in
the vertical irregularity that characterizes the BEJ. For this reason, the vehicle is modeled in detail,
detail, while for the BEJ, a simple geometry, able to describe the unevenness, was considered.
while for the BEJ, a simple geometry, able to describe the unevenness, was considered. In addition,
In addition, only the concrete slab was modeled to reduce the computational burden, its bottom
only the concrete slab was modeled to reduce the computational burden, its bottom surface being
surface being fully constrained.
fully constrained.
Appl. Sci. 2017, 7, 599 8 of 20
Appl. Sci. 2017,
Appl. Sci. 7, 599
2017, 7, 8 of 20
21
Appl. Sci. 2017, 7, 599 8 of 20

Figure 6. LS-DYNA 3D model.

Figure 6.
Figure 6. LS-DYNA
LS-DYNA 3D
3D model.
model.
The pavement is a three-layer (two asphalt concrete and one cement concrete) 22.4 m long and
The
4 m-wide pavement
The pavement
pavement
model. The is aa three-layer
is BEJ is located(two
three-layer (two
at theasphalt
middleconcrete
asphalt concrete and one
and (Figure
of the strip cement
one cement
cement concrete)
concrete)
7), where 22.4
22.4
there22.4 m
m
is am long and
longhole
0.8-m and
The is a three-layer (two asphalt concrete and one concrete) long and
4
4
in m-wide
m-wide
the lowermodel.
model.
layer,The
The BEJ
BEJ
while is
is
the located
located
two at
at
upper the
the middle
middle
layers of
of
are the
the strip
strip
interrupted(Figure
(Figure
by the 7),
7), where
where
BEJ (2.00 there
there
m is
is a
a
long). 0.8-m
0.8-m
The hole
hole
joint
4 m-wide model. The BEJ is located at the middle of the strip (Figure 7), where there is a 0.8-m hole in
in
in the
the lower
lower layer,
layer, while
while the
the two
the two upper
upper layers
layers are interrupted
arethe
interrupted by the BEJ
BEJm(2.00
havem long). The joint
surface
the is 3layer,
lower cm lower
while than
the two rolling
upper surface,
layers arethus upper
interrupted thebyBEJthe
bypavement (2.00
layers
(2.00 long).maThe
long). The
transition
joint joint
zone
surface
surface
surface is
is 3 cm
m 3before
cm lower
lower than the rolling surface, thus the upper pavement layers have a transition zone
of
is 31.60
cm lower than thethan
and afterthethe
rolling rolling
surface, surface,
joint thus
to reach thus
thethe
the upper upper pavement
appropriate
pavement height.
layers havelayers
The have a transition
pavement/joint
a transition zone
zone of model
1.60 m
of
of 1.60
1.60
consists m
mof before
before
15540 and
and
8-nodeafter
after the
the
brick joint
joint
(solid)to
to reach
reach
elements.the
the appropriate
appropriate
Mechanical height.
height. The
The
characteristics pavement/joint
pavement/joint
of the materialsmodel
model
are
before and after the joint to reach the appropriate height. The pavement/joint model consists of 15,540
consists of
of 15540
consistsbrick
linear-elastic 15540
for the8-node
8-node
bottom brick
brick
layer (solid)
(solid)
(concreteelements.
elements. Mechanical
Mechanical
slab-structure) and characteristics
characteristics
viscoelastic of
of the materials
the upper
materials are
are
8-node (solid) elements. Mechanical characteristics of the materialsfor arethe two
linear-elastic layers.
for the
linear-elastic
linear-elastic for the bottom
for the bottom layer (concrete
layer (concrete slab-structure)
slab-structure) and
and viscoelastic
viscoelastic for the two upper layers.
for the two upper layers.
bottom layer (concrete slab-structure) and viscoelastic for the two upper layers.

Figure 7. Pavement and BEJ model (detail of the BEJ zone).

Figure BEJ zone).
Figure 7.
7. Pavement
Pavement and
and BEJ
BEJ model
model (detail
(detail of
of the
the BEJ zone).
With respect to the boundary conditions adopted in the model, the horizontal displacements on
With
the sides respect
With of
respect to
to the
to
the model theboundary
the boundary
boundary
were conditions
conditions
conditions
restrained adopted
adoptedin
adopted
to represent in
thethe
in the
the model,
model,
model,
confinement the horizontal
thethe to thedisplacements
horizontal
horizontal
due displacements
displacements
presence on
on
of the
the
on sides
the of
sides the
of model
the
the sides of portion
surrounding model
the model were
were restrained
restrained totorepresent
represent the
the confinement
confinement due
due
were restrained to represent the confinement due to the presence
of pavement. to
to the
the presence
presence of
of the
the
surrounding
surrounding portion
The test portion
vehicle ofof pavement.
pavement.
model (Figure 8) was initially built to perform roadside hardware crash test
The
The test
test vehicle
vehicle model
simulations [35–37]. It is the (Figure
model (Figure
IVECO 8) 8) was180
was
FIAT initially
initially built to
built
NC, IVECO to(Group
performCNH
perform roadside
roadside hardware
hardware
Industrial), crashItaly,
crash
Torino, test
test
simulations
simulations
simulations
4-axle [35–37].
heavy[35–37].
[35–37]. ItIt
Itisis
goods vehicle, the
isthe
the IVECO
IVECO
IVECO
with a net FIAT
FIAT
FIAT
mass 180
180of NC,
180 NC,
NC,
10,560 IVECO
IVECO
IVECO
kg and a(Group
(Group CNH
(Group
total CNH
CNH
mass Industrial),
Industrial),
Industrial),
of about 30,000 Torino,
Torino, Italy,
Torino,
kg. Italy,
4-axle
Italy,
Its model
4-axle
heavy heavy
goods
4-axle heavy
consists goods
goods
of about vehicle,
vehicle, with
vehicle,
12,000 with
a net
with
shell, a net
150amass mass
net mass
solid and of
of 10,560
of 10,560
10010,560
beam kg
kg and
kg and
and
elements a total
a total
a total
(in mass
mass
mass
total, of about
ofabout
about
of about 30,000
30,000
30,000
15,000 kg.
kg. Its
kg.Its model
Its model
elements). model
consists of about
consists of about 12,000
12,000 shell,
shell, 150
150 solid
solid and
and 100
and 100 beam
100 beam elements
beam elements
elements (in(in total,
(in total, about
total, about 15,000
about 15,000 elements).
15,000 elements).
elements).

Figure 8. FE model of the 30-ton truck.

Figure 8. FE model of the 30-ton truck.
Figure
Figure 8.
8. FE
FE model
model of
of the
the 30-ton
30-ton truck.
truck.
Appl. Sci. 2017, 7, 599 9 of 21
Appl.
Appl.Sci.
Sci.2017,
2017,7,7,599
599 99of
of20
20

The
The whole
whole suspension
suspensionsystem
systemof ofthe
thevehicle
vehicletest
testwas
wasmodeled
modeledto toreproduce
reproduce the
the dynamic
dynamiceffects
effects
and the vehicle behavior during
behavior during
and the vehicle behavior the test, as shown
during the test, as shown in Figure
shown in Figure 9.
9. The reference truck has an
Figure 9. The reference truck has an old
oldtype
typeofof
suspension
suspension(leaf
(leafsprings),
springs),with
withelastic
elasticsteel
steel beams
beamspacked
packed over
over each
eachaxle,
axle,and
andhydraulic
hydraulic dampers.
dampers.
(leaf springs), with elastic steel beams packed over each axle, and hydraulic dampers.

Figure
Figure9.
9.Detail
Detailof
ofthe
therear
rearsuspension
suspensionsystem.
suspension system.
system.

These
Theseelements
elementswere
werereproduced
reproducedwith withaaavariable
variablethickness
thicknessshell
shellbeam,
beam,while
whilethetheaxles
axlesare
aresteel
steel
These elements were reproduced with variable thickness shell beam, while the axles are steel
beams. The
beams. The main parts of the vehicle’s structure have an elastic-plastic behavior. The loads
loads are
beams. The main
main parts
parts of the vehicle’s
of the vehicle’s structure
structure have
have anan elastic-plastic
elastic-plastic behavior.
behavior. TheThe loads are
are
accurately
accurately transferred
transferred to
to the
the pavement
pavement because
because this
this vehicle
vehicle is
is equipped
equipped with
with inflated
inflated tires
tires and
and
accurately transferred to the pavement because this vehicle is equipped with inflated tires and rolling
rolling
rolling wheels,
wheels, as shown
as in
shown in Figure 10.
10. AAsurface
contact surface with
with aa Coulomb-like friction ensures the
wheels, as shown Figurein10.
Figure
A contact contactwith
surface
a Coulomb-likeCoulomb-like frictionthe
friction ensures ensures the
interface
interface between
interface the
between the two
the two bodies (tire-rubber
bodies (tire-rubber and pavement).
and pavement).
between two bodies (tire-rubber and pavement).

(a)
(a) (b)
(b)
Figure
Figure 10.
10.Pneumatic
10. Pneumatictire
Pneumatic tirecomponents
componentsadopted
adoptedin
in(a)
(a)the
thevehicle
vehiclemodel
modeland
and(b)
(b)the
thetire
tire model
modelfront
model front
front
section during a general
during aa general
section during time
general time step.
time step.
step.

The
Themodel
modelwaswasvalidated
validatedandandthen
thenused
usedto tocalculate
calculatethe
thestress,
stress,strain
strainand
andacceleration
accelerationof ofeach
each
The model was validated and then used to calculate the stress, strain and acceleration of each
element of
element of the
of the pavement,
the pavement, joint
pavement, joint and
joint and vehicle.
and vehicle. These
vehicle. These data
These data are
data are useful,
are useful, for
useful, for example,
for example, to
example, to verify
to verify the
verify the fatigue
the fatigue
fatigue
element
damage
damage of
of the
thevehicle
vehiclestructure,
structure, estimating
estimating the
the effects
effectsof
of an
anuneven
uneven surface
surface as
asthe
the BEJ
BEJ examined.
examined.
damage of the vehicle structure, estimating the effects of an uneven surface as the BEJ examined.
3.4.
3.4.Front
Frontand
andRear
RearAxle
AxleQuarter
QuarterCar
CarModel
Model
3.4. Front and Rear Axle Quarter Car Model
The
Thelinear
linearquarter-truck
quarter-truckmodel
model(Figure
(Figure11a)
11a)described
describedin in[32]
[32]was
wasimplemented
implementedin inMATLAB
MATLABto to
The linear quarter-truck model (Figure 11a) described in [32] was implemented in MATLAB to
estimate
estimate the
the dynamic
dynamic load
load increment
increment induced
induced by
by the
the BEJ
BEJ at
at several
several speeds.
speeds. The
The basic
basic equations
equations of
of
estimate the dynamic load increment induced by the BEJ at several speeds. The basic equations of
motion
motionforforthe
theexamined
examinedmodel
modelareare(1)
(1)and
and(2):
(2):
motion for the examined model are (1) and (2):
.. . .
xu ==ccs ((xxs −−xxu ))−−((kks ++kkt ))xxu ++kks xxs ++kkt u,u ,
u x
mm
muu xuu = css ( xss − xuu ) − (kss + kt t ) xuu + kss xss + kt t u ,
(1)
(1)
(1)
Appl. Sci. 2017, 7, 599 10 of 20

Appl. Sci. 2017, 7, 599

ms xs = cs ( xu − xs ) + ks ( xu − xs ) . 10 of 21
(2)

where mu is the unsprung mass, ..ms is the . sprung . mass, k is the suspension spring constant,
ms x s = cs ( x u − x s ) + k s ( xu − xs s ). (2)
kt is the tire spring constant, cs is the suspension viscous damping constant, xu is the vertical
where mu is the unsprung mass, ms is the sprung mass, k s is the suspension spring constant, k t is the
displacement
tire of the unsprung
spring constant, mass, xs viscous
cs is the suspension is the verticaldamping displacement
constant, xof the
u is thesprung mass,
vertical xs is the
displacement
.
of the unsprung mass, xs is the vertical displacement of the sprung mass, x s is the time derivative of
time derivative of vertical displacement of the . sprung mass in m, xu is the time derivative of
vertical displacement of the sprung mass in m, x u is the time derivative of vertical displacement of the
unsprung
..
vertical displacement
mass, x s is the of acceleration
the unsprung mass,
of the sprung xs ismass, ..
the acceleration of the sprung
x u is the acceleration mass, xu is
of the unsprung the
mass
and u is the road
acceleration of theprofile mass and u is the road profile elevations.
elevations.
unsprung
This
This mathematical
mathematical model model was
was formulated
formulated in in the
the form
form of of state
state equations.
equations. TheThe general
general form
form of
of the
the
state
state model
model for
for aa linear
linear dynamic
dynamic system
system isis reported
reported in in Equations
Equations(3) (3)and
and(4):
(4):

x x==Ax
Ax++Bu,
Bu ,
. (3)
(3)

==Cx
yy Cx++Du,
Du, (4)
(4)

where AAisisthe
where thestate
state matrix,
matrix, B the
B the inputinput matrix,
matrix, C theCoutput
the output
matrix,matrix, D the transmission
D the direct direct transmission
matrix,
xmatrix, x is the state vector and u and y are, respectively, the input (profile elevations) and the
is the state vector and u and y are, respectively, the input (profile elevations) and the output (reaction
output vectors
forces) (reaction forces)11b).
(Figure vectors (Figure 11b).

V  0 1 0 0 
mS xs − ks − cs ks cs 
 ms ms ms ms 
A=
0 0 0 1 
 ks − (k s + kt ) − cs 
ks cs
cs
 m mu mu mu 
 u

 0  C = [0 0 kt 0]
mu xu

 0 
D = [− kt ]
B =  0 
kt
u
x = [xs xs xu xu ]
 kt 
u = [u ]
y
 mu 
(a) (b)
Figure 11.
Figure 11. Scheme
Scheme of
of (a)
(a) the
thequarter-truck
quarter-truck model
model and
and (b)
(b) the
thematrices
matrices used
used in
inthe
thestate-space
state-space model
model
in MATLAB.
in MATLAB.

The mechanical parameters of the front and rear axles used in the model are consistent with the
The mechanical parameters of the front and rear axles used in the model are consistent with the
data provided in [30] (Table 3).
data provided in [30] (Table 3).
Table 3. Mechanical parameters of the quarter-truck model [32].
Table 3. Mechanical parameters of the quarter-truck model [32].
Parameter Value
Parameter 1350Value
kg (front)
ms (sprung mass)
1470
1350 kgkg(front)
(rear)
ms (sprung mass)
335 kg (front)
1470 kg (rear)
mu (unsprung mass)
450kg
335 kg(front)
(rear)
mu (unsprung mass) 270 kN/m (front)
ks (suspension spring constant) 450 kg (rear)
660 kN/m (rear, steel)
270 kN/m (front)
ks (suspension springconstant)
kt (tire spring constant) 760660
kN/m (standard radial)
kN/m (rear, steel)
16.8 kN s/m (front)
cs (suspension viscous
kt (tire spring damping constant)
constant) 760 kN/m (standard radial)
89.0 kN s/m (rear)
16.8 kN s/m (front)
cs (suspension viscous damping constant)
89.0 kN s/m (rear)
Appl. Sci. 2017, 7, 599 11 of 21

Appl. Sci. 2017, 7, 599 11 of 20

According
Appl. to Equation (5), the dynamic load coefficient (DLC) was calculated to characterize
Sci. 2017, 7, 599 the
11 of 20
According to Equation (5), the dynamic load coefficient (DLC) was calculated to characterize
magnitude of dynamic tire forces [9]:
the magnitude
Accordingoftodynamic tire
Equation forces
(5), [9]:
the dynamic load coefficient (DLC) was calculated to characterize
the magnitude of dynamic tire forces [9]: RMS dynamic tire force
DLC = RMS dynamic tire force (5)
DLC = Static tire force (5)
Static
RMS tire force
dynamic tire force
where RMS is the root mean squareDLC of the=dynamic tire force signal calculated along the road profile, (5)
often indicated using the σ notation. of
where RMS is the root mean square ThetheRMS
dynamicStatic
tire tiretire
force (σ) force
force signal
can calculatedby
be calculated along theof
means road profile,(6):
Equation
often
whereindicated using
RMS is the themean
root σ notation.
squareThe RMS
of the tire force
dynamic (σ)force
tire can be calculated
signal by means
calculated alongofthe
Equation (6):
road profile,
ZT
often indicated using the σ notation. The 1 Ttire force (σ) 1canNNbe calculated by means of Equation (6):
RMS
σσ2 2== 1 ff((tt))22dtdt== 1 ∑ f kf2 ,2 ,
T10 
T T N
N1k=k =1N1
k  (6)
(6)

σ = 0 f (t ) 2 dt =  f k2 ,
2
(6)
where f is the dynamic tire force f(t) at timeT t0= k × ∆t, ∆tNis k
the =1 simulation output time step and
where fk is the dynamic tire force f(t) at time t = × Δt, Δt is the simulation output time step and T is T is
k
the
thetotal
totalfsimulation
where simulationtime
time(T N×
(T== N ∆t).
× Δt).
k is the dynamic tire force f(t) at time t = k × Δt, Δt is the simulation output time step and T is

the total simulation time (T = N × Δt).

4.4.Test
TestResults
Resultsand andDiscussion
Discussion
4. Test
TheResults
The and Discussion measurements
geometric-topographic
geometric-topographic permittedplotting
measurements permitted plottingthe
thelevel
levelcurves
curvesofof
thethe surveyed
surveyed
pavement
The area,
pavement area, which
which isis characterized
characterized
geometric-topographic by the
by the presence
measurements presence ofof
permitted threeBEJs
three
plotting BEJslevel
the marked
marked asasBEJ1,
curves ofBEJ1, BEJ2
the BEJ2 andand
surveyed
BEJ3
BEJ3 (Figure 12). In particular, three different longitudinal alignments, highlighted
pavement area, which is characterized by the presence of three BEJs marked as BEJ1, BEJ2 and12,
(Figure 12). In particular, three different longitudinal alignments, highlighted in Figure
in 12,
Figure
were
were
BEJ3 examined.
examined.
(Figure 12). In particular, three different longitudinal alignments, highlighted in Figure 12,
were examined.
BEJ1 BEJ2 BEJ3
Alignment 1 BEJ1 BEJ2 BEJ3
Alignment 1 Alignment 2
Alignment 3
Alignment 2
Alignment 3

Figure 12. Level curves (in m) of the surveyed road.

Figure 12. Level curves (in m) of the surveyed road.
Figure 12. Level curves (in m) of the surveyed road.
In Figure 13, the location of all three examined BEJs along the alignments is depicted. The most
critical
InIn BEJ,
Figureclose
Figure 13, to
13, which
the many
location
the location accidents
ofof
all threewere
allthree localized,
examined
examined BEJs
BEJsis along
BEJ1,
alongwhich is in a sag is
the alignments
the alignments section (Figure
isdepicted.
depicted. 13).
The
The most
most
critical
criticalBEJ,
BEJ,close
closetotowhich
which many accidentswere
many accidents werelocalized,
localized,is is BEJ1,
BEJ1, which
which isainsag
is in a sag section
section (Figure
(Figure 13). 13).
Alignment 1 Alignment 2 Alignment 3
BEJ1 BEJ2 BEJ3
52.0 Alignment 1 Alignment 2 Alignment 3
BEJ1 BEJ2 BEJ3
52.0
51.5

51.5
(m) (m)

51.0
Elevation

51.0
Elevation

50.5

50.5
50.0

50.0
49.5
0 50 100 150 200 250 300 350 400 450 500
49.5
0 50 100 150 200Distance
250 (m) 300 350 400 450 500
Distance (m)
Figure 13. Road longitudinal profiles and the BEJs’ location.
Figure 13. Road longitudinal profiles and the BEJs’ location.
Figure 13. Road longitudinal profiles and the BEJs’ location.
Appl. Sci. 2017, 7, 599 12 of 21
Appl. Sci. 2017, 7, 599 12 of 20

In the
the examined
examinedportion
portionofof
thethe
road,
road,a rotation of the
a rotation of carriageway
the carriagewaytakes takes
place, place,
and BEJ1
andisBEJ1
placedis
in correspondence
placed with thewith
in correspondence transition curve, while
the transition curve,the other
while thetwo (i.e.,two
other BEJ2 and
(i.e., BEJand
BEJ2 3) are
BEJlocated
3) are
in the circle
located in thearc. Applying
circle a high-pass
arc. Applying Butterworth
a high-pass filterfilter
Butterworth withwith
cut-off spatial
cut-off frequency
spatial frequency equal to
equal
0.05 cycles/m
to 0.05 cycles/m(wavelength
(wavelengthofof2020m)m)totothe
thethree
three500
500m-long
m-longalignments,
alignments,ititwaswas found
found that
that the most
significant
significant irregularities
irregularitieswere
wereplaced
placedinincorrespondence
correspondencetotothe thethree BEJs
three BEJs compared
compared to to
thethe
elevations in
elevations
the remaining
in the remaining part of the
part pavement
of the pavement profiles (Figure
profiles 14).14).
(Figure

BEJ1 BEJ2 BEJ3

0.03

0.02

0.01
Elevation (m)

0.00
0 100 200 300 400 500

-0.01

-0.02

-0.03
Distance (m)

Figure 14. Filtered longitudinal profile: Alignment 2.

Moreover,Figure
Moreover, Figure1414 shows
shows thatthat the unevenness
the unevenness caused caused
by theby the expansion
expansion joint
joint in the in transition
road the road
transition curve (i.e., BEJ1) is much greater than the other two BEJs. For this
curve (i.e., BEJ1) is much greater than the other two BEJs. For this reason, further analyses reason, further analyses
were
were specifically carried out for BEJ1
specifically carried out for BEJ1 (Figure 13). (Figure 13).
From the
From the inspection
inspection survey,
survey, thethe largest
largest pavement
pavement deformation
deformation waswas found
found 25 m after
25 m after the
the joint
joint
location. Other pavement unevenness characterized by wavelengths of about
location. Other pavement unevenness characterized by wavelengths of about 15 m occurred after 15 m occurred after the
BEJ1.
the They
BEJ1. Theyare are
linked to the
linked dynamic
to the dynamic load
loadapplied
appliedtotovehicles
vehiclesbumping
bumpingon onthe
thejoint
jointwith
with aa speed of
speed of
90–110 km/h and the principal oscillation frequency of about 2
90–110 km/h and the principal oscillation frequency of about 2 Hz (body bounce).Hz (body bounce).
The roughness
The roughness assessment
assessment of the survey
of the survey road
road section
section was
was performed
performed calculating
calculating thethe most
most
worldwide used
worldwide used roughness
roughness evaluation
evaluation criteria
criteria [38]:
[38]: thethe International
International Roughness
Roughness IndexIndex (IRI).
(IRI).
This approach [39] is based on a mathematical model (quarter-car), whose mechanical
This approach [39] is based on a mathematical model (quarter-car), whose mechanical parameters (also parameters
(also known
known as golden-car)
as golden-car) are listed
are listed in Tablein Table
4. 4.

Table 4. Golden-car
Table 4. Golden-car mechanical
mechanical parameters
parameters for
for the
the IRI
IRI quarter-car
quarter-car model.
model.

Parameter Value
Parameter
ms (sprung mass) Value
9000 kg
mmus (unsprung
(sprung mass)
mass) 9000
1350kg
kg
mu (unsprung
ks (suspension mass)
spring constant) 1350kN/m
569.7 kg
ks (suspension spring
kt (tire spring constant)
constant) 569.7
5877kN/m
kN/m
kt (tire viscous
cs (suspension spring constant)
damping constant) 5877 kN/m
54 kN s/m
cs (suspension viscous damping constant) 54 kN s/m

Its calculation is performed using Equation (7):

Its calculation is performed using Equation (7):
l/v
1
IRI = Zl/v x s − xu ⋅ dt (7)
1 l 0 . . 
IRI = x s − x u · dt
 (7)
l
0
Appl. Sci. 2017, 7, 599 13 of 21

Appl. Sci. 2017, 7, 599 .

13 of 20
where l is the length of the profile in km, v is the simulated speed equal to 80 km/h, x s is the time
.
derivative of vertical displacement of the sprung mass in m and x u is the time derivative of vertical
where l is the length of the profile in km, v is the simulated speed equal to 80 km/h, xs is the time
displacement of the unsprung mass in m. The final value is expressed in slope units (e.g., m/km
derivative of vertical displacement of the sprung mass in m and xu is the time derivative of vertical
or mm/m).
displacement
IRI was developed of the unsprung
in order tomass in m.
assess notTheonlyfinalthe value is expressed
ride quality on road in slope units (e.g.,
pavements, but m/km
also other
or mm/m).
detrimental effects, such as dynamic load increment (on both vehicle and pavement) induced by road
pavementIRI was developed
irregularities. Thein order to assess
algorithm to benotused
only forthe its
ridecalculation
quality on road pavements,
is provided but also other
in [40].
detrimental effects, such as dynamic load increment (on both vehicle and pavement) induced by
The Italian highway road agency requires calculating IRI on a segment of 10 m, admitting in
road pavement irregularities. The algorithm to be used for its calculation is provided in [40].
general values lower than 1.8 m/km [38]. In the case of short span bridges and when joints are present
The Italian highway road agency requires calculating IRI on a segment of 10 m, admitting in
on road pavement
general surface,
values lower thanIRI 1.8values
m/km until
[38]. In2.5the
m/km case ofare tolerated
short span bridges[41]. According to Figure
and when joints are 15,
all examined
present on alignments
road pavementhave IRI valuesIRI
surface, greater
valuesthan the2.5threshold
until m/km are limit suggested
tolerated [41]. by [41] in the
According to area
across BEJ1,15,
Figure while close toalignments
all examined the other havetwo BEJs (i.e., greater
IRI values BEJ2 andthanBEJ3) and in the
the threshold limitremaining
suggested by part[41]of the
profiles, thearea
in the IRIacross
valuesBEJ1,
are lower thantothe
while close thethreshold
other twovalue. TheBEJ2
BEJs (i.e., only andexception
BEJ3) andisinAlignment
the remaining 2 across
BEJ3 (IRI = 3.55 m/km). Considering the significant difference found between the three BEJs inisterms
part of the profiles, the IRI values are lower than the threshold value. The only exception
of IRIAlignment
values, it 2isacross
evidentBEJ3 (IRIBEJ1
that = 3.55has
m/km). Considering
the most critical the significant difference found between the
unevenness.
three BEJsininsitu
Several termsmeasurements
of IRI values, it is evident
were that BEJ1
carried out has
to the most critical
evaluate unevenness.
the dynamic consequences on
Several in situ measurements were carried out to evaluate the dynamic consequences on heavy
heavy vehicles running on BEJ1. All of the recorded accelerometer signals were made uniform to
vehicles running on BEJ1. All of the recorded accelerometer signals were made uniform to a
a homogeneous time duration of T = 4.587 s and considering a time sampling equal to 5.6 × 10−4 s (∆t).
homogeneous time duration of T = 4.587 s and considering a time sampling equal to 5.6 × 10-4 s (Δt).
Thus,Thus,
a wholea wholenumbernumber of Nof =N 8192
= 8192samples
sampleswas was registered
registered for foreach
eachsignal.
signal. In addition,
In addition, forofall
for all theof the
vertical acceleration measures, the reference
vertical acceleration measures, the reference time time t 0 corresponds to test vehicle transit
t0 corresponds to test vehicle transit on a on a transversal
road transversal
section located at about
road section 29 m at
located before
aboutthe29 m BEJ.
before the BEJ.

Alignment 1 Alignment 2 Alignment 3 IRI threshold

BEJ1 BEJ2 BEJ3
12

10

8
IRI (m/km)

0
0 50 100 150 200 250 300 350 400 450 500
Distance (m)

Figure
Figure 15. International
15. International roughnessindex
roughness index (IRI)
(IRI) values
valuescalculated
calculatedonon
a segment of 10ofm.
a segment 10 m.

Table 5 shows the time of the three axles’ passages on the joint for each loading test, captured by
Table
means5ofshows
a highthe time
frame ofcamera.
rate the three axles’ passages on the joint for each loading test, captured by
means of a high frame rate camera.
Table 5. Time of bump between axles and joint.
5. Time
Table Test of bumpt1 between
Number (s) axlestand
t2 (s) 3 (s)
joint.
1 0.999 1.184 1.230
Test Number 2 t1 (s)
1.273 t2 (s)
1.509 1.567 t3 (s)
3 1.038 1.231 1.278
1 0.999 1.184 1.230
4 1.258 1.491 1.549
2 1.273 1.509 1.567
3 1.038 1.231 1.278
4 1.258 1.491 1.549
Appl. Sci. 2017, 7, 599 14 of 21

Appl. Sci. 2017, 7, 599 14 of 20

Appl. Sci. 16
Figure 2017, 7, 599
shows the recorded accelerometer signals provided by the accelerometers on14the
of 20front,
Figure 16 shows the recorded accelerometer signals provided by the accelerometers on the
medium and rear axles (number of tests = 1), with the indication of the three bump axle times. Positive
front,Figure
medium16 and rearthe
shows axles (number
recorded of tests = 1), with
accelerometer the provided
signals indicationbyof the three bump axleon
accelerometers times.
the
peak Positive
values peak
higher
front, medium
than
values
and rear
1axles
g were
higher than 1found
g were
(number
for
found
of tests
all ofwith
for
= 1),
the
all test
ofthe vehicle’s
theindication
test axles,
vehicle’s axles,
of the
confirming
threeconfirming the
bump axlethe
loss of
loss
times.
tire/pavement
of
Positive contact
tire/pavement due
contact
peak values to the
duethan
higher passage
to the on
1 gpassage BEJ1.
on BEJ1.
were found for all of the test vehicle’s axles, confirming the loss
of tire/pavement contact due to the passage on BEJ1.
Test Number 1
Test Number 1
Front Axle Medium Axle Rear Axle
Front Axle t1 Medium
t2 t3 Axle Rear Axle
2
t1 t2 t3
2
1.5
1.5
(g) (g)

1
Acceleration

1
Acceleration

0.5
0.5
Vertical

0
Vertical

0
-0.5
-0.5
-1
-1 0 0.5 1 1.5 2 2.5
0 0.5 1 Time (s) 1.5 2 2.5
Time (s)
Figure 16. An example of accelerometer signals (front, medium and rear axles: Test Number 1) and
Figure 16. An example of accelerometer signals (front, medium and rear axles: Test Number 1) and
axles’
Figurebumps
16. Antransiting
example ofover the bridge expansion
accelerometer joint.medium and rear axles: Test Number 1) and
signals (front,
axles’ bumps transiting over the bridge expansion joint.
axles’ bumps transiting over the bridge expansion joint.
Analyzing the maximum positive vertical accelerations obtained for each configuration test
Analyzing
(Table 2), it the
Analyzing maximum
wasthe
found that positive
maximum thepositive
mediumvertical accelerations
axle was
vertical obtained
the most obtained
accelerations critical forfor
one each
(Figure
each configuration
17). test test
In addition,
configuration
Figure
(Table 17
2), highlights
it was the
found importance
that the of the
medium moving
axle speed:
was the increasing
most the
critical vehicle
one velocity
(Figure 17).
(Table 2), it was found that the medium axle was the most critical one (Figure 17). In addition, Figureresults
In in an 17
addition,
increment
Figure 17 of vertical
highlights acceleration
the importance on vehicles
of the and
moving passengers.
speed: increasing the vehicle velocity
highlights the importance of the moving speed: increasing the vehicle velocity results in an increment results in an
increment of vertical acceleration on vehicles
of vertical acceleration on vehicles and passengers. and passengers.

Front Axle Medium Axle Rear Axle

Front Axle Medium Axle Rear Axle
2
value

1.82
value
peak

1.8
1.6
peak

1.6
1.4
- positive

1.4
1.2
- positive

1.2
1
(g) (g)

0.81
acceleration

0.8
acceleration

0.6
0.6
0.4
0.4
0.2
Vertical

0.2
0
Vertical

0 1 (V=107 km/h) 2 (V=84 km/h) 3 (V=103 km/h) 4 (V=85 km/h)

1 (V=107 km/h) Test Number
2 (V=84 km/h) 3 (V=103 km/h) 4 (V=85 km/h)
Test Number

Figure 17. Maximum positive vertical acceleration values recorded during each test.
Figure Maximum
17. 17.
Figure positive
Maximum positivevertical
vertical acceleration valuesrecorded
acceleration values recorded during
during eacheach
test.test.
Appl. Sci. 2017, 7, 599 15 of 21
Appl. Sci. 2017, 7, 599 15 of 20

All
Allofofthe
thecollected
collected data
data were
were also
also analyzed in the
analyzed in the frequency
frequencydomain,
domain,and andthethe results
results were
were
represented
representedusing
usingthe
theacceleration
acceleration power
power spectral density(PSD)
spectral density (PSD)function.
function.This
Thisprocedure
procedure permitted
permitted
studyingthe
studying theprincipal
principaloscillation
oscillation and
andvibration
vibrationcomponents
components of of
thethe
vehicle duedue
vehicle to the
to bump (Figure
the bump 18). 18).
(Figure

0.02
Body bounce Pitching
(~2 Hz) (~3 Hz) Body roll
0.018 front axle
medium axle
0.016 rear axle

Body bounce
0.014

0.012 Body bounce

(~2 Hz)
PSD (G /Hz)

Pitching
2

0.01

0.008

Axle hop
0.006

Body roll
0.004 Axle hop
(~1 Hz)
(~10-20 Hz)
Tyre oscillation
0.002 Vibration
(~40-60 Hz)
(>100 Hz)

0
0.1 1 10 100 1000
Frequency (Hz)

Figure18.18.Power
Figure Powerspectral
spectraldensity
density(for
(foreach
eachaxle,
axle,average
averagevalue
valueofofthe
thefour
fourtests
tests with
with thethe
20 g
20 g accelerometer).
accelerometer).

AccordingtotoFigure
According Figure18,18,the
themain
main oscillation
oscillation components
componentsare: are:
• • Vehicle
Vehicle roll oscillation: body roll f ∼≅ 1.0 Hz (T∼≅ 1 s) (sprung mass);
roll oscillation: body roll f = 1.0 Hz (T = 1 s) (sprung mass);
• Primary vehicle oscillation: body bounce f ∼ ≅ 2.0 Hz (T ≅ 0.5 s) (sprung mass);
• Primary vehicle oscillation: body bounce f = 2.0 Hz (T ∼ = 0.5 s) (sprung mass);
• Vehicle pitching oscillation: pitching f ≅ 3.0 Hz (T ≅ 0.33 s) (sprung mass);
• • Vehicle f ∼ ∼
Primary vehicle axles oscillation: axle hop f ≅ 10–20 Hz (Ts) ≅(sprung
pitching oscillation: pitching = 3.0 Hz (T = 0.33 0.1–0.05mass);
s) (unsprung mass);
• • Primary vehicle axles oscillation: axle hop f ∼ 10–20 Hz (T ∼ 0.1–0.05 s) (unsprung
=
Tire oscillation and vibration: f ≅ 40–60 Hz (T ≅ 0.025–0.017 s) (unsprung = mass); mass);
• • Tire oscillation and vibration:
Power and other part in vibration: f ∼ 40–60 Hz (T ∼ 0.025–0.017
= f > 100 Hz (T=< 0.001 s) (vibration).s) (unsprung mass);
• Power and other part in vibration: f > 100 Hz (T < 0.001 s) (vibration).
Lower frequency responses (1–4 Hz) characterize the vehicle’s body (sprung mass) bounces, or
pitches
Lower andfrequency
rolls. High-frequency
responses (1–4 responses (10–50 Hz), instead,
Hz) characterize correspond
the vehicle’s bodyto(sprung
axle hopmass)
vibrations.
bounces,
Acceleration
or pitches measurements and
and rolls. High-frequency their (10–50
responses processing permitted
Hz), instead, identifying
correspond all hop
to axle vibrational
vibrations.
components
Accelerationinduced in the vehicle
measurements andtest byprocessing
their the geometric irregularity
permitted of BEJ1.
identifying allInvibrational
particular,components
the most
induced in the vehicle test by the geometric irregularity of BEJ1. In particular, the most importantthe
important ones, which have a frequency range 1–4 Hz, determine the complete lifting from ones,
pavement
which have of the rear axles
a frequency of the
range 1–4running ordinarythe
Hz, determine heavy vehicles.
complete Consequently,
lifting this phenomenon
from the pavement of the rear
reduces
axles of thethe adherence
running of wheel-pavement,
ordinary heavy vehicles. until the completethis
Consequently, lossphenomenon
of adherence.reduces
Furthermore, in the
the adherence
paved areas adjacent to the joint, more accentuated irregularities appear, due to the considerable
of wheel-pavement, until the complete loss of adherence. Furthermore, in the paved areas adjacent to
dynamic increment of the loads transmitted to the pavement.
the joint, more accentuated irregularities appear, due to the considerable dynamic increment of the
In this sense, several simulation analyses were performed using the linear quarter-truck model
loads transmitted to the pavement.
described in [32] to estimate the dynamic load increment for the heavy vehicles’ characteristic speed
In this sense, several simulation analyses were performed using the linear quarter-truck model
range running on the examined road (i.e., 80–120 km/h). The results of the simulation analyses
described in [32] to estimate the dynamic load increment for the heavy vehicles’ characteristic speed
(Figure 19) show higher DLC values, increasing vehicle speed in analogy with literature
range
studies [34,37].onThe
running theDLCexamined roadwas
calculation (i.e.,performed
80–120 km/h). The results
considering of the segment
just a limited simulation analyses
having a
(Figure 19) show higher DLC values, increasing vehicle
total length of 100 m: 50 m before and 50 m after the BEJ1. speed in analogy with literature studies [34,37].
The DLC calculation
Comparing thewas
DLCperformed
obtained for considering
two different justalignments
a limited segment
(i.e., 1 andhaving
3), theahighest
total length
valuesofwere
100 m:
50found
m before and 50 m after the BEJ1.
for Alignment 3, whose unevenness (due to the BEJ presence) was greater than Alignment 1.
Comparing the DLC obtained for two different alignments (i.e., 1 and 3), the highest values were
found for Alignment 3, whose unevenness (due to the BEJ presence) was greater than Alignment 1.
Appl. Sci. 2017, 7, 599 16 of 21
Appl. Sci. 2017, 7, 599 16 of 20

In Appl.
fact,
In
Sci. 2017,
thethe
fact,
7, 599 of the carriageway along the transition curve has its torsion axis (inner roadside)
rotation
rotation
16 of 20
of the carriageway along the transition curve has its torsion axis (inner roadside)
very close
very closeto to
Alignment
Alignment1.1.
In fact, the rotation of the carriageway along the transition curve has its torsion axis (inner roadside)
very close to Alignment 1.
front-mean value rear-mean value front-alignment1 front-alignment 3 rear-alignment 1 rear-alignment 3
0.5
front-mean value rear-mean value front-alignment1 front-alignment 3 rear-alignment 1 rear-alignment 3
0.5
0.4

0.4
0.3
DLC DLC

0.3
0.2

0.2
0.1

0.1
0
70 80 90 100 110 120 130
0 Speed (km/h)
70 80 90 100 110 120 130
Speed (km/h)
Figure 19. Front and rear axles dynamic load coefficient (DLC) values at different speeds.
Figure 19. Front and rear axles dynamic load coefficient (DLC) values at different speeds.
Figure 19. Front and rear axles dynamic load coefficient (DLC) values at different speeds.
Figure 19 also reports the DLC mean value for both front and rear axles calculated considering
allFigure 19 also reports
three alignments. the DLC
A major mean value
dispersion for both
(referred to thefront
DLC and rearvalue)
mean axles calculated
was foundconsidering
for the rearall
Figure 19 also reports the DLC mean value for both front and rear axles calculated considering
three alignments.
axle, A major
probably caused dispersion
by the (referred
mechanical parametersto theconsidered.
DLC mean value) was found for the rear axle,
all three alignments. A major dispersion (referred to the DLC mean value) was found for the rear
probably caused
The resultsbyobtained
the mechanical
from theparameters
DLC analysis, considered.
in analogy to what was found through in situ
axle, probably caused by the mechanical parameters considered.
The results
acceleration obtained from
measurements, the DLC that
highlighted analysis,
the rearin analogy
axle is the to most
whatcritical
was found
one inthrough
terms ofinthesitu
The results obtained from the DLC analysis, in analogy to what was found through in situ
acceleration
dynamic load measurements,
transmitted tohighlighted
the pavement that
andthe rearbridge
to the axle is the most critical one in terms of the
structure.
acceleration measurements, highlighted that the rear axle is the most critical one in terms of the
dynamic DLC values
load were also
transmitted tocalculated for the
the pavement andfirst
to 100
the m of thestructure.
bridge examined road branch, where no BEJs
dynamic load transmitted to the pavement and to the bridge structure.
areDLC
present,
valuesto better
were underline the effects caused100 by BEJ1. Figure 20 shows thatbranch,
the DLC of thenofirst
DLC values werealso
alsocalculated
calculatedfor
for the
the first
first 100 m m ofof the
the examined
examined road
road where
branch, where no BEJsBEJs
100
areare m-long
present, road
toto has
better values significantly lower than those calculated in the area surrounding BEJ1.
present, betterunderline
underlinethe theeffects
effects caused
caused by BEJ1. Figure
by BEJ1. Figure20 20shows
showsthatthatthethe DLC
DLC of of
thethe first
first
100100
m-long road has values significantly lower than those calculated in the area
m-long road has values significantly lower than those calculated in the area surrounding BEJ1. surrounding BEJ1.
Rear axle

First 100 m Rear axle

100 m surrounding the BEJ

0.5 First 100 m 100 m surrounding the BEJ

0.5
0.4

0.4
0.3
DLC DLC

0.3
0.2

0.2
0.1

0.10
70 80 90 100 110 120 130
0
Speed (km/h)
70 80 90 100 110 120 130
Speed (km/h)
Figure 20. Dynamic load coefficient (DLC) values at different speeds before the bridge expansion
joint and across it (Alignment 3).
Figure
Figure 20.20. Dynamic
Dynamic load
load coefficient
coefficient (DLC)
(DLC) values
values at different
at different speeds
speeds before
before thethe bridge
bridge expansion
expansion joint
joint and across it (Alignment 3).
and across itthe
Finally, (Alignment 3).
stress distribution on heavy vehicle mechanical components due to the transit on
BEJ1 was investigated by means of a 3D model. The obtained results are depicted in Figure 21: since
Finally, the stress distribution on heavy vehicle mechanical components due to the transit on
theFinally,
maximum depth
the stress of the measured
distribution onaheavyelevation
vehicle is about 2.5 cm, the stress magnitude is not
BEJ1 was investigated by means of 3D model. Themechanical components
obtained results due to
are depicted inthe transit
Figure 21: on BEJ1
since
particularly
was burdensome for vehicle components.
theinvestigated
maximum by means
depth of ameasured
of the 3D model.elevation
The obtained results
is about 2.5 are
cm,depicted in magnitude
the stress Figure 21: since
is notthe
particularly burdensome for vehicle components.
 Appl.
Appl. Sci.Sci. 2017,
2017, 7, 7,
599599 17 of
1720
of 21

Thus, the location of a BEJ in a transition curve does not provide a significant increase in the
mechanical
maximum depth components of road vehicles,
of the measured elevationwhile it meaningfully
is about 2.5 cm, theaffects
stress traffic safetyis(due
magnitude not to the loss
particularly
of tire/pavement
burdensome contact)
for vehicle and dynamic load increment (i.e., higher forces transmitted to the road
components.
pavement and bridge structure).

(a)

(b)

(c)
Figure
Figure 21.21. Heavy
Heavy vehiclestress
vehicle stress(MPa)
(MPa)atat90
90km/h:
km/h: (a)
(a) full
fullvehicle,
vehicle,(b)
(b)magnification
magnificationofof
medium and
medium and
rear axles and (c) magnification of front axles.
rear axles and (c) magnification of front axles.
Appl. Sci. 2017, 7, 599 18 of 21

Thus, the location of a BEJ in a transition curve does not provide a significant increase in the
mechanical components of road vehicles, while it meaningfully affects traffic safety (due to the loss
of tire/pavement contact) and dynamic load increment (i.e., higher forces transmitted to the road
pavement and bridge structure).
For these reasons, the road agency planned:

• the reconstruction of the joint to permit a more regular and continuous variation of the planimetric
geometry of the road platform;
• measures for the reduction of the driving speed (i.e., speed cameras and speed enforcement) to
ensure a better circulation.

The adopted strategies ensure that under ordinary conditions, the circulation is regular, and the
accident risk for road users is reduced; moreover, the reduced number and frequency of the required
maintenance works increase the traffic safety.

5. Conclusions
In road engineering, bridge expansion joints often cause troubles concerning evenness and skid
resistance. Even if the joint is designed and built in accordance with the best practice, the discontinuity
of the pavement may produce a dangerous unevenness. This irregularity could be very important
if the joint is in a transition section (clothoid) of the road, as considered in the paper. In fact, in this
section, the curvature and super-elevation of the road gradually change between two consequent
elements with different planimetric curvature. Consequently, in this area, the road surface is “slanted”,
and if the joint is large (for example in a highway seismic joint), the elevation difference from the points
of the road can reach some centimeters. This difference could produce an additional, undesirable
unevenness in addition to a possible construction unevenness.
A heavy vehicle was instrumented with piezometric accelerometers to evaluate the effects of
the irregular BEJ on a running vehicle. The results of the performed surveying and tests highlighted
the unevenness on the surface profile caused by the presence of the BEJ along a transition curve.
The analyses carried out in the present work underline how this kind of discontinuity may affect the
response of truck suspensions and the dynamic response of the bridge. The measurements permitted
also identifying the most representative frequencies of loads applied to the structure. The results also
confirmed the possibility of a complete lifting of the vehicle’s rear axles with serious risks of grip
loss on the road and, consequently, serious detrimental effects on road traffic safety. The road agency
adopted two main actions to improve the level of service in the road section and to ensure a safe and
efficient traffic flow: reduction of geometric irregularities resulting from the joint and reduction of the
allowed speed.
BEJs are localized irregularities that can have different severity levels according to many factors,
such as if placed in the road transition curve. This study provides a guidance on the evaluation of
the detrimental effects induced by BEJs, which may affect all of the road infrastructure components:
vehicles, pavement, bridge and users, with attention to safety. The results can be used by road agencies
to specify strategies and procedures for policies to guarantee minimum safety levels at roads.

Author Contributions: Giuseppe Loprencipe had the original idea for the study. Giuseppe Loprencipe and
Paola Di Mascio carried out all in situ measurements. Giuseppe Loprencipe, Laura Moretti and Pablo Zoccali
analyzed the data. Lorenzo Puzzo implemented the finite element model in LS-DYNA and performed the
data post-processing analysis. Giuseppe Loprencipe and Pablo Zoccali wrote sections for the first draft of the
manuscript. All authors contributed to further drafts and had full access to all of the data.
Conflicts of Interest: The authors declare no conflict of interest.
Appl. Sci. 2017, 7, 599 19 of 21

Abbreviations
The following abbreviations are used in this manuscript:

BEJ Bridge expansion joint

DIVINE Dynamic Interaction between Vehicles and Infrastructure Experiment
DLC Dynamic load coefficient
IRI International Roughness Index
PSD Power spectral density
RMS Root mean square

References
1. Corazza, M.V. Uno sguardo sul ponte. L’industria italiana del Cemento 2007, 836, 718–731.
2. Lachinger, S.; Hoffmann, S. Level Crossing. 2015. Available online: http://www.mageba.ch/data/docs/en_
CH/4839/NEWS-level-crossing-strengthening-rips-Issue79.pdf (accessed on 5 April 2017).
3. Lima, J.M.; de Brito, J. Inspection survey of 150 expansion joints in road bridges. Eng. Struct. 2009, 31,
1077–1084. [CrossRef]
4. Pesterev, A.V.; Bergman, L.A.; Tan, C.A. Pothole-induced contact forces in a simple vehicle model.
J. Sound Vib. 2002, 256, 565–572. [CrossRef]
5. Palamas, J.; Coussy, O.; Bamberger, Y. Effects of surface irregularities upon the dynamic response of bridges
under suspended moving loads. J. Sound Vib. 1985, 99, 235–245. [CrossRef]
6. Loprencipe, G.; Zoccali, P. Ride Quality Due to Road Surface Irregularities: Comparison of Different Methods
Applied on a Set of Real Road Profiles. Coatings 2017, 7, 59. [CrossRef]
7. Dexter, R.J.; Connor, R.J.; Kaczinski, M.R. NCHRP Report 402; Transportation Research Board, National
Research Council: Washington, DC, USA, 1997; Available online: http://onlinepubs.trb.org/Onlinepubs/
nchrp/nchrp_rpt_402.pdf (accessed on 27 April 2017).
8. Eurocode 1 EN 1991-2. Actions on Structures. Part 2: Traffic Loads on Bridges; European Committee for
Standardisation: Brussels, Belgium, 2003.
9. Cebon, D. Handbook of Vehicle-Road Interactions; Swets & Zeitlinger: Lisse, The Netherlands, 1999.
10. Green, M.F.; Cebon, D. Dynamic interaction between heavy vehicles and highway bridges. Comput. Struct.
1997, 62, 253–264. [CrossRef]
11. Au, F.T.K.; Cheng, Y.S.; Cheung, Y.K. Effects of random road surface roughness and long-term deflection
of prestressed concrete girder and cable-stayed bridges on impact due to moving vehicles. Comput. Struct.
2001, 79, 853–872. [CrossRef]
12. Kim, C.W.; Kawatani, M.; Kim, K.B. Three-dimensional dynamic analysis for bridge–vehicle interaction with
roadway roughness. Comput. Struct. 2005, 83, 1627–1645. [CrossRef]
13. Cantieni, R. Investigation of vehicle-bridge interaction for highway bridges. In Heavy Vehicles and Roads
Technology, Safety and Policy; Thomas Telford Publishing: London, UK, 1992; pp. 130–137.
14. Cantieni, R.; Krebs, W.; Heywood, R. OECD IR6 DIVINE Project, Dynamic Interaction between Vehicle and
Infrastructure Experiment, Element 6, Bridge Research; EMPA Test Report No. 153’031; EMPA: Duebendorf,
Switzerland, 2000.
15. Chatti, K.; Lee, D. Development of new profile-based truck dynamic load index. J. Transp. Res. Board 2002,
1806, 149–159. [CrossRef]
16. Yang, Y.B.; Lin, C.W. Vehicle–bridge interaction dynamics and potential applications. J. Sound Vib. 2005, 284,
205–226. [CrossRef]
17. Tan, G.H.; Brameld, G.H.; Thambiratnam, D.P. Development of an analytical model for treating bridge-vehicle
interaction. Eng. Struct. 1998, 20, 54–61. [CrossRef]
18. Loprencipe, G.; Zoccali, P. Use of generated artificial road profiles in road roughness evaluation. JMT 2017,
25, 24–33. [CrossRef]
19. Willis, J. Correlation of calculated and measured dynamic behaviour of bridges. In Proceedings of the
Symposium on Dynamic Behaviour of Bridges at the Transport and Road Research Laboratory, Crowthorne,
UK, 19 May 1977.
Appl. Sci. 2017, 7, 599 20 of 21

20. Leonard, D.R.; Eyre, R. Damping and Frequency Measurements on Eight Box Girder Bridges; TRRL Laboratory
Report 682; Transport and Road Research Laboratory: Crowthome, UK, 1975.
21. Eyre, R.; Tilly, G.P. Damping measurements on steel and composite bridges. In Proceedings of the Symposium
on Dynamic Behaviour of Bridges at the Transport and Road Research Laboratory, Crowthorne, UK,
19 May 1977.
22. Woodrooffe, J.H.; Le Blanc, P.A.; Le Piane, K.R. Effects of suspension variations on the dynamic wheel loads
of a heavy articulated highway vehicle. In Vehicle Weights and Dimensions Study, Technical Report Volume II;
Transportation Association of Canada: Ottawa, ON, Canada, 1986.
23. Bonin, G.; Cantisani, G.; Loprencipe, G.; Ranzo, A. Effetti dinamici nei giunti delle pavimentazioni
aeroportuali in calcestruzzo. Industria Italiana del Cemento 2007, 77, 590–607.
24. Moretti, L.; Di Mascio, P.; D’Andrea, A. Environmental impact assessment of road asphalt pavements.
Mod. Appl. Sci. 2013, 7, 1. [CrossRef]
25. Cantisani, G.; Loprencipe, G. Road roughness and whole body vibration: Evaluation tools and comfort
limits. J. Transp. Eng. 2010, 136, 818–826. [CrossRef]
26. Lima, J.M.; de Brito, J. Management system for expansion joints of road bridges. Struct. Infrastruct. Eng.
2010, 6, 703–714. [CrossRef]
27. Loprencipe, G.; Cantisani, G.; di Mascio, P. Global assessment method of road distresses. In Life-Cycle
of Structural Systems: Design, Assessment, Maintenance and Management, Proceedings of the 4th International
Symposium on Life-Cycle Civil Engineering, Tokyo, Japan, 16 November 2014; CRC Press: Leiden, The Netherlands,
2015; Available online: https://www.crcpress.com/Life-Cycle-of-Structural-Systems-Design-Assessment-
Maintenance-and-Management/Furuta-Frangopol-Akiyama/p/book/9781138001206 (accessed on
28 February 2017).
28. Bonin, G.; Folino, N.; Loprencipe, G.; Oliverio, R.C.; Polizzotti, S.; Teltayev, B. Development of
a Road Asset Management System in Kazakhstan. In Proceedings of the TIS 2017 International
Congress on Transport—Infrastructure and Systems, Rome, Italy, 10–12 April 2017; Available online:
https://www.crcpress.com/Transport-Infrastructure-and-Systems-Proceedings-of-the-AIIT-International/
DellAcqua-Wegman/p/book/9781138030091. (accessed on 28 February 2017).
29. Loprencipe, G.; Pantuso, A.; di Mascio, P. Sustainable Pavement Management System in Urban Areas
Considering the Vehicle Operating Costs. Sustainability 2017, 9, 453. [CrossRef]
30. Sayers, M.W.; Karamihas, S.M. The Little Book of Profiling; The Regent of the University of Michigan: Ann
Arbor, MI, USA, 1998.
31. Steyn, W.J.V. Moving Dynamic Loads Caused by Bridge Deck Joint Unevenness: A Case Study.
2004. Available online: https://www.researchgate.net/profile/Wynand_Steyn/publication/30510498_
Guidelines_for_incorporation_of_vehiclepavement_interaction_effects_in_pavement_design/links/
02e7e531bc9450f72e000000.pdf (accessed on 17 January 2017).
32. Streit, D.A.; Kulakowski, B.T.; Sebaaly, P.E.; Wollyung, R.J. Road Simulator Study of Heavy-Vehicle Wheel Forces
No. FHWA-RD-98–019; Research and Development Turner-Fairbank Highway Research Center: Georgetown
Pike McLean, VA, USA, 1998.
33. Istituto Nazionale di Statistica (ISTAT). Rapporto Incidenti Stradali Anno 2015. 2016. Available online:
www.istat.it (accessed on 15 September 2016). (In Italian)
34. Lu, Y.; Yang, S.; Li, S.; Chen, L. Numerical and experimental investigation on stochastic dynamic load of
a heavy duty vehicle. Appl. Math. Model. 2010, 34, 2698–2710. [CrossRef]
35. Bonin, G.; Cantisani, G.; Loprencipe, G.; Ranzo, A. Improvement of portable concrete barrier design using
computational mechanics. J. Transp. Res. Board 2006, 28, 3–13. [CrossRef]
36. Bonin, G.; Cantisani, G.; Ranzo, A.; Loprencipe, G.; Atahan, A.O. Retrofit of an existing Italian bridge rail for
H4a containment level using simulation. Int. J. Heavy Veh. Syst. 2009, 16, 258–270. [CrossRef]
37. Múčka, P. Road Roughness Limit Values Based on Measured Vehicle Vibration. J. Infrastruct. Syst. 2016, 23.
[CrossRef]
38. Múčka, P. International Roughness Index specifications around the world. Road Mater. Pavement 2016, 18,
1–37. [CrossRef]
39. Sayers, M.W. On the calculation of international roughness index from longitudinal road profile.
Transp. Res. Rec. 1995, 1501, 1–12.
Appl. Sci. 2017, 7, 599 21 of 21

40. American Society of Testing and Materials. Standard Practice for Computing International Roughness Index
of Roads from Longitudinal Profile Measurements; ASTM E1926:08 International: West Conshohocken, PA,
USA, 2015.
41. ANAS S.p.a. Bozza Capitolato Speciale d’Appalto 2010. Available online: http://www.tecnometer.
it/images/stories/normative/Bozza_Capitolato_ANAS_Pavimentazioni_Stradali_2010.pdf (accessed on
27 April 2017).

© 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/).