Seismic assessment and retrofitting measures of a historic stone masonry bridge

Article (PDF Available)inEarthquakes and Structures 10(3):645-667 · March 2016with 227 Reads
DOI: 10.12989/eas.2016.10.3.645
Abstract
The 750 m long “De Bosset” bridge in the Cephalonia Island of Western Greece, being the area with the highest seismicity in Europe, was constructed in 1830 by successive stone arches and stiff block type piers. The bridge suffered extensive damages during past earthquakes, such as the strong M7.2 earthquake of 1953, followed by poorly-designed reconstruction schemes with reinforced concrete. In 2005, a multidisciplinary project for the seismic assessment and restoration of the “De Bosset” bridge was undertaken under the auspices of the Greek Ministry of Culture. The proposed retrofitting scheme combining soil improvement, structural strengthening and reconstruction of the deteriorated masonry sections was recently applied on site. Design of the rehabilitation measures and assessment of the pre- and post-interventions seismic response of the bridge were based on detailed in-situ and laboratory tests, providing foundation soil and structural material properties. In-situ inspection of the rehabilitated bridge following the strong M6.1 and M6.0 Cephalonia earthquakes of January 26th and February 3rd 2014, respectively, revealed no damages or visible defects. The efficiency of the bridge retrofitting is also proved by a preliminary performance analysis of the bridge under the recorded ground motion induced by the above earthquakes.
Earthquakes and Structures, Vol. 10, No. 3 (2016) 645-667
DOI: http://dx.doi.org/10.12989/eas.2016.10.3.645 645
Copyright © 2016 Techno-Press, Ltd.
http://www.techno-press.com/journals/eas&subpage=7 ISSN: 2092-7614 (Print), 2092-7622 (Online)
Seismic assessment and retrofitting measures of a historic
stone masonry bridge
Emmanouil N. Rovithis1 and Kyriazis D. Pitilakis2
1Institute of Engineering Seismology and Earthquake Engineering (EPPO-ITSAK), Dasyliou Str., Elaiwnes,
Pylaia, 55102, Thessaloniki, Greece
2Department of Civil Engineering, Aristotle University of Thessaloniki, 54124, Thessaloniki, Greece
(Received June 23, 2015, Revised January 19, 2016, Accepted January 20, 2016)
Abstract. The 750 m long “De Bosset” bridge in the Cephalonia Island of Western Greece, being the area
with the highest seismicity in Europe, was constructed in 1830 by successive stone arches and stiff block-
type piers. The bridge suffered extensive damages during past earthquakes, such as the strong M7.2
earthquake of 1953, followed by poorly-designed reconstruction schemes with reinforced concrete. In 2005,
a multidisciplinary project for the seismic assessment and restoration of the “De Bosset” bridge was
undertaken under the auspices of the Greek Ministry of Culture. The proposed retrofitting scheme
combining soil improvement, structural strengthening and reconstruction of the deteriorated masonry
sections was recently applied on site. Design of the rehabilitation measures and assessment of the pre- and
post-interventions seismic response of the bridge were based on detailed in-situ and laboratory tests,
providing foundation soil and structural material properties. In-situ inspection of the rehabilitated bridge
following the strong M6.1 and M6.0 Cephalonia earthquakes of January 26th and February 3rd 2014,
respectively, revealed no damages or visible defects. The efficiency of the bridge retrofitting is also proved
by a preliminary performance analysis of the bridge under the recorded ground motion induced by the above
earthquakes.
Keywords: retrofitting; “De Bosset” bridge; old stone masonry bridges; numerical analysis;
micropiles; Cephalonia 2014 earthquakes
1. Introduction
The problem of retrofitting a two-hundred years old stone masonry bridge having sustained
serious earthquake-induced damages followed by poorly-designed restorations requires
consideration of a large number of parameters. The latter are related to the foundation soil
conditions, the heterogeneity of construction materials, the existing strength and pathology of the
structure and the identification of possible failure modes within assessment of the structure’s
vulnerability. The problem becomes more complex and challenging if the structure is exposed to
high seismic risk, large traffic loads and unfavorable environmental conditions, while founded
through rigid massive foundations on deformable soil, thus introducing soil-structure interaction as
Corresponding author, Researcher Ph.D., E-mail: rovithis@itsak.gr
Emmanouil N. Rovithis and Kyriazis D. Pitilakis
a key mechanism in controlling the dynamic characteristics of the structure and the effective
seismic motion (Page 1993, Barros and Luco 1995, Proske and Van Gelder 2009, Asteris et al.
2014, Preciado et al. 2015). In this regard, a successful rehabilitation scheme should balance
between structural integrity upgrade aligned to design codes requirements and maintenance of the
bridge monumental features. For this reason, a comprehensive set of available in-situ inspections,
field measurements and earthquake recordings are of major importance to assess the pathology and
select the most appropriate retrofitting scheme for such type of monumental structures (Krstevska
et al. 2008, Cakir and Seker 2015).
A well-documented case study dealing with the design of an innovative retrofitting scheme to
meet both structural safety requirements and architectural needs of the historic stone masonry “De
Bosset” bridge in the Cephalonia Island of Western Greece is presented in this paper. The pre-
rehabilitation seismic stability of the bridge in terms of transverse structural strength and bearing
capacity of the foundation soil is estimated by a fracture line approach and a stress-based seismic
analysis of a three-dimensional finite-element model, respectively. Foundation soil properties are
defined by field surveys and laboratory tests. The design seismic motion at the foundation level of
the bridge is determined from soil response analysis accounting for non-linear soil behavior under
selected ground motions recorded in close distance from the bridge site at the time of the study.
Soil compliance is modeled by means of Winkler spring supports to evaluate soil-structure
interaction effects by comparing the seismic response between fixed- and flexible-base (compliant)
bridge models in the realm of elastodynamic considerations. Based on the identified pathology and
the critical zones of the bridge, a set of intervention measures is proposed by combining
foundation soil improvement, strengthening of the bridge stability under transverse loading and
maintenance of its monumental nature. The proposed rehabilitation works were recently
completed, providing an overall seismic upgrade of the bridge. The latter is proved by the very
good seismic performance of the rehabilitated bridge following the two strong M6.1 and M6.0
earthquakes of January 26th and February 3rd, 2014, respectively, that stroke Cephalonia inducing
particularly high ground accelerations. Besides post-earthquake field inspection, further support on
the effectiveness of the intervention scheme is provided from numerical analyses of the
rehabilitated bridge by implementing the actual seismic loading induced by the two earthquakes.
The study presented herein was performed in the framework of a multidisciplinary project
under the auspices of the Directorate for the Restoration of Byzantine and Post-byzantine
Monuments of the Greek Ministry of Culture (Pitilakis 2006, Rovithis and Pitilakis 2011).
2. Historical background
Originally constructed in 1830 by successive stone arches founded on stiff block-type stone
piers, the historic multi-span stone masonry “De Bosset” bridge connects the shorelines of
Argostoli and Drapano at the southern side of the Argostoli bay (Fig. 1). The bridge has a total
length of 750 m and its height varies between 2 to 4 m along its longitudinal axis. The foundation
area of the bridge piers is approximately equal to 10 m long and 5 m wide. The severe Ms7.2
earthquake of 1953 (Papazachos 1997) induced extensive deformations of the deck and differential
settlement of the piers (Fig. 2(a)). The characteristic failure mechanism of old stone bridges related
to out-of-plane collapse of the arch walls and filling material (Griffith et al. 2003) is evident in
Fig. 2(b). Major parts of the bridge were totally reconstructed during the period of 1960 to 1970.
The original stone sections and the filling material of the bridge were removed (Fig. 2(c)). A new
646
Seismic assessment and retrofitting measures of a historic stone masonry bridge
Fig. 1 Overview of the old stone masonry 750 m long “De Bosset” bridge in Cephalonia island connecting
the city of Argostoli with the opposite shoreline (Drapanos) (Google earth image)
reinforcement mesh was installed to form the arch-shaped sections of the bridge (Fig. 2(d)) and
two new concrete spandrel walls of approximately 40 cm width were cast in place (Fig. 2(e)). In
this manner, the original architectural pattern of the bridge was reproduced by reinforced concrete
(Fig. 2(f)) while modifying substantially the material homogeneity and structural stiffness of the
bridge. However, the extent of the above interventions along the longitudinal axis of the bridge is
unknown since the authentic dry stone material was preserved in some sections of bridge including
the first two arches close to the city of Argostoli. The latter is illustrated in the photogrammetric
mapping of a part of the western façade of the bridge shown in Fig. 3.
3. Pathology of the bridge
In 2005, a multidisciplinary research project for the seismic assessment and restoration of the
“De Bosset” bridge was undertaken by the Laboratory of Soil Mechanics, Foundation and
Geotechnical Earthquake Engineering (LSMFGEE) of the Aristotle University of Thessaloniki
(Pitilakis 2006) and supervised by the Directorate for the Restoration of Byzantine and Post-
byzantine Monuments of the Greek Ministry of Culture. A series of in-situ inspections and
laboratory tests on concrete and stone specimens was performed to obtain a detailed knowledge of
the bridge pathology and strength. With reference to structural defects, extensive cracks up to 20
cm wide and 40 cm deep were recorded in the tensile zones under the arches (Fig. 4(a)). An
underwater inspection revealed large foundation scouring up to 1m deep in most of the bridge
647
Emmanouil N. Rovithis and Kyriazis D. Pitilakis
piers (Fig. 4(b)), due to strong sea currents within the Argostoli bay. The original stone masonry
material was highly deteriorated by deep corrosion, loss of mass and poor condition of joint
mortars. Delamination and detachment of covering was observed for the concrete sections.
Laboratory tests on concrete specimens revealed a substantially low compressive strength
corresponding to concrete category C8/10, whereas the effective cross section of the reinforcement
bars was reduced to almost one half of the original section area. Mean values for the static
modulus of elasticity were measured at 10 GPa. Further details on the above experimental data
may be found in Papayianni et al. (2007).
Fig. 2 (a), (b) Extensive damages to the “De Bosset” bridge due to the severe Ms7.2 earthquake of 1953,
including lateral deformations and differential settlement of the deck and out-of-
p
lane collapse of the arch
walls and filling material; (c)-(f) partial reconstruction of the bridge with reinforced concrete during the
period 1960-1970; (c) removal of the original stone sections and the filling material; (d), (e) placement o
f
the reinforcement forming the original shape of the bridge arches (f) a typical arch of the bridge
reconstructed with reinforced concrete
648
Seismic assessment and retrofitting measures of a historic stone masonry bridge
1234
567
Original stone material Reiforced concrete
Original stone material
Original stone materialReiforced concrete
Sea level
Fig. 3 Western façade of the bridge along the first seven arches close to Argostoli city based on
hotogrammetric mapping (The photogrammetric study was conducted in 2002 by the Photogrammetry Lab
of the School of Rural and Surveying Engineering of the National Technical University of Athens)
Fig. 4 Typical findings during the in-situ inspection of the bridge pathology in 2005: (a) Extensive
longitudinal and transverse cracks in the tensile zones under the arches and (b) foundation scouring of the
piers base
0
5
10
15
20
25
30
35
0 20 40 60 80 100 120
Borehole B1
Borehole B2
SPT blows
z (m
)
(b)
Fig. 5 (a) Geotechnical and geophysical tests performed along the “De Bosset” bridge in 2005, (b) SPT test
results obtained at B1 and B2 boreholes location
649
Emmanouil N. Rovithis and Kyriazis D. Pitilakis
4. Geotechnical and geophysical investigation
The geotechnical and geophysical field campaign comprised of geotechnical borings, Standard
Penetration Testing (SPT), Cross-Hole Seismic (CH) and Microtremor Array Measurements
(MAM) at selected locations along the bridge shown in Fig. 5(a). SPT profiles at boreholes B1 and
B2 are plotted in Fig. 5(b). The field surveys were complemented with laboratory tests to provide
the mechanical and the dynamic properties of the foundation soil. The geotechnical cross section
of the bridge site obtained from the in-situ surveys is illustrated in Fig. 6. The local sandstone,
which can be considered as the seismic bedrock with shear wave propagation velocity (Vs) larger
than 1000 m/s, is found at a depth of about 35 m. It is overlaid by a soft silty clay layer of low
shear strength with Vs varying between 140 and 170 m/s. The measured angle of friction (φu) and
cohesion (cu) of the surficial layer under undrained loading conditions are 3o and 26.1 KPa,
respectively, while the average plasticity index (PI) is 30%. Considering the above characteristics,
the foundation soil of the bridge should be classified as soil type D (i.e., soft clays of high
plasticity index and thickness higher than 10 m) according to the Greek Seismic Code (EAK2000).
Resonant column tests provided shear modulus (G) degradation and hysteretic damping ratio (D)
curves of the soil layers with increasing shear strain (γ). The latter were adopted in equivalent
linear soil response analyses, presented in the ensuing.
5. Design input motion and fundamental frequency of soil
In order to estimate the design seismic motion at the foundation level of the bridge, a well-
documented set of twenty five earthquakes with magnitude (Ms) and peak horizontal ground
acceleration (PHGA) ranging between 3.7 to 5.2 and 0.02 g to 0.2 g, respectively, were selected to
estimate the rock outcrop motion. The above earthquakes were recorded in the period of 1999 to
2003 by an accelerometric station, hereafter referred as CH station, installed at the Cephalonia
Greek Telecommunication building (OTE), 500m away from the bridge site. The CH station was
operated and maintained by ITSAK (www.itsak.gr) as part of the Greek National Accelerometric
Network until 2012, when the CH station was replaced by a new 24 bit accelerometric station
(ARG2) installed at the basement of a two-story building, owned by the Prefecture authority of the
Ionian Islands (EPPO-ITSAK report 2014a). Further details on the ARG2 station are reported in
the ensuing subsoil conditions at the CH station site are available from earlier studies (AUTH-
ITSAK report 1996) based on SPT and CH tests.
Upon implementing the shear wave velocity profile at CH station site, a series of deconvolution
analyses were performed to derive the rock outcrop motion for each one of the selected earthquake
records. The computed rock motions were scaled to 0.36 g, corresponding to the design
acceleration for the seismic zone III which includes the island of Cephalonia according to the
Greek Seismic Code EAK2000. The scaled rock motions were then specified at the seismic
bedrock (i.e., Vs>1000 m/s) defined at -35 m of the bridge soil profile. One-dimensional equivalent
linear ground response analyses were then performed by implementing the finite-element code
Cyberquake (Modaressi and Foerster 2000) to compute soil response at the ground surface of the
bridge site. For this reason, the experimental G-γ-D curves mentioned above were employed.
Based on the geotechnical cross section derived from the in-situ geophysical tests (Fig. 6), three
shear wave velocity profiles grouped as “A1-A4”, “A5-A9” and “A10-A15” (Fig. 7(a)) were
adopted to model the variation of the foundation soil stiffness along the bridge longitudinal axis.
650
Seismic assessment and retrofitting measures of a historic stone masonry bridge
Fig. 6 Geotechnical cross section of the bridge foundation soil along the line B1-B2 shown in Fig. 5(a)
0
5
10
15
20
25
30
35
0 100 200 300 400 500 600
V
s
(m/sec)
z (m)
A1 - A4
A5 - A9
A10 - A15 (a)
0
1
2
3
4
5
9901
9902
9903
9904
9905
9906
2000-01
2000-02
2000-03
2000-04
2000-05
2001-01
2001-02
2001-03
2001-04
2001-05
2001-06
2003-01
2003-02
2003-03
2003-05
2015
6040
6042
1858
A1-A4
A5-A9
A10-A15
PGA
s
/ PGA
r
25 earthquake re cordings
(b)
Fig. 7 (a) Shear wave propagation velocity (Vs) profiles to account for variation of Vs along the bridge
longitudinal axis. (b) Surface-to-bedrock peak acceleration ratios (PGAs/PGAr) at the bridge site for each
one of the 25 selected earthquake recordings
651
Emmanouil N. Rovithis and Kyriazis D. Pitilakis
0
1
2
3
4
5
6
00.511.522.53
25 earthquake recordings
Mean spectrum
Series30
T
(
sec
)
S
a
/ PGA
Mean spectrum ± σ
(a)
0
1
2
3
4
5
6
00.511.522.53
25 earthquake recordings
Mean spectrum
Series30
EAK 2000 (Soil D)
EC8 (Soil C)
T
(
sec
)
S
a
/ PGA
Mean spectrum ± σ
(b)
0
1
2
3
4
5
6
7
8
012345678910
Analytical solution
(linear analysis)
25 earthquake recordings
(EQL analysis)
f(Hz)
Surface-to-Rock Outcrop
Trasnfer Function
(c)
Fig. 8 Normalized acceleration response spectra at (a) bedrock level and (b) ground surface of the bridge soil
rofile computed from equivalent linear soil response analysis. EAK2000 and EC8 normalized elastic
spectrum corresponding to soil category D and C, respectively, are also plotted. (c) Comparison of surface-
to-rock outcrop transfer functions between damped (d=10%) linear elastic solution (Roesset 1977) an
d
equivalent linear analyses. The graphs refer to the “A1-A4” V
s
model
652
Seismic assessment and retrofitting measures of a historic stone masonry bridge
The corresponding surface-to-bedrock peak acceleration ratios (PGAs/PGAr) are plotted in Fig.
7(b). Each code name in the abscissa of the graph refers to a selected earthquake event. A mean
PGAs/PGAr ratio at 1.7 was revealed, corresponding to a mean peak horizontal acceleration of 0.6
g at the ground surface of the bridge soil profile. It is noted that, stronger amplification of the
seismic motion from bedrock to ground surface was observed for the deeper soil models “A5-A9”
and “A10-A15”. Figs. 8(a) and 8(b) shows the normalized acceleration response spectra computed
at the bedrock and the ground surface, respectively, referring to the “A1-A4” shear wave velocity
model. The high frequency content of the computed earthquake motions is reflected in the mean
response spectrum compared with the elastic design spectrum of EAK2000 and EC8 (CEN 2002)
for soil category D and C, respectively. Surface-to-rock outcrop transfer functions obtained from
the equivalent linear analyses are compared in Fig. 8(c). The analytical solution reported in
Roesset (1976) for a multi-layer damped (d=10%) soil is also plotted, denoting a fundamental
frequency (fo) of the bridge soil deposit at 4 Hz under linear viscoelastic considerations. Evidently,
the degradation of soil stiffness and damping increase with increasing shear strains γ shifted fo to
lower frequencies in the range of 2.5 to 3.5 Hz, depending on the frequency characteristics of the
input motion.
6. Transverse structural stability
Mention has already been made to the extensive longitudinal and transversal cracks that were
observed under the stone arches, indicating poor structural stability. The fracture line method
reported in Erdogmus and Boothby, 2004 was employed to evaluate the transverse strength of the
bridge under seismic loading. The above method is originated from the yield line theory that has
been adopted in code provisions (British Standards 1992) for unreinforced masonry design. In this
case, the spandrel wall is considered as a retaining wall of the material fill volume above the arch,
inducing active-state pressures as applied loads. In fracture line analysis, the yield lines on a
surface, representing axes of rotation, are the equivalent of hinges on a beam that occur when the
plastic bending moment is reached (Erdogmus and Boothby 2004). The bending moment
producing rupture is assigned to all points along a hinge line. The selected fracture line pattern
representing the failure mechanism, depends on the boundary conditions, geometry and material
properties of the problem at hand. In the case of the “De Bosset” bridge, the assumed fracture line
pattern is shown in Fig. 9, resembling the out-of-plane collapse of the spandrel walls caused by the
1953 earthquake (Figs. 2(a) and 2(b)). According to the above method, the maximum allowable
pressure qmax, referring to the transverse strength of the wall, may be estimated from the
equivalence of external and internal work
ie
W= W
(1)
where Wi is the internal work done by the resisting wall, given by
=
fi
WMlθ (2)
and We stands for the external work done by the applied loads as follows
=
e
WQθd (3)
653
Emmanouil N. Rovithis and Kyriazis D. Pitilakis
Hinge 1
(l=2.12m) Hinge 5
(l=2.12m)
Hinge 3
(l=4.20m)
Hinge 2
(l=1.50m) H=1.50m
Hinge 4
(l=1.50m)
45
o
45
o
AB
C
0.40m
P
AE
δ
H/3
Fig. 9 Assumed fracture line pattern of a typical bridge arch for estimating transverse load effects by
means of the fracture line method reported in Erdogmus and Boothby (2004). The adopted fracture
line pattern resembles that of the out-of-plane collapse pattern of the “De Bosset” induced by the 1953
earthquake
Table 1 Simplified assessment of the bridge transverse stability based on the fracture line method reported in
Erdogmus and Boothby (2004)
Internal Work calculations
Panel Hinge l (m) θ Μf=frS
(KNm/m) Wi=Mflθ (ΚΝm)
A 1 2.12
θ2 4.32 13.81θ
2 1.5 θ 4.32 6.91 θ
B 3 4.2 θ 4.32 18.14 θ
C 4 1.5 θ 4.32 6.91 θ
5 2.12
θ2 4.32 13.81 θ
ΣWi= 57.01θ
External Work calculations
Panel l (m) H (m) θ Q (KN) d (m) We=Qdθ (ΚΝm)
A 1.5 1.5
θ2
max H
ql6
0.375qmax
H2
0.75
0.397 qmax θ
B 4.2 1.5 θ
max H
ql2
3.15qma
x
H3
0.5
1.58 qmax θ
C 1.5 1.5
θ2
max H
ql6
0.375qmax
H2
0.75 0.397 qmax θ
ΣWe= 2.37 qmax θ
Maximum allowable pressure qmax (KPa)=ΣWi/ΣWe 24.05
Active-state earth pressures (Mononobe-Okabe method)
H (m) kh/g
kv/g φ(o)/δ(ο) ψ(ο) KAE P
AE(KN/m) qAE
(KPa)
1.5 0.6
0.25 37/18.5 38.7 1.738 25.2 33.6
Safety factor against shear failure SFs=qmax/qAE 0.71
654
Seismic assessment and retrofitting measures of a historic stone masonry bridge
In the above equations, Mf (=frS) is the bending moment producing fracture, fr is the tensile
strength of the stone masonry, S stands for the section modulus of the wall, l is the length of the
fracture line, θ is the rotation along the fracture line, Q refers to the resultant force imposed by the
external loads and d is the distance from the point of application of Q to the facture line.
Upon substituting We and Wi in Eq. (1), it may be rewritten as
=
r
fSlθQθd (4)
With reference to the external seismic loading imposed on the bridge spandrel wall, the
resultant lateral active thrust PAE (Fig. 9) is estimated by implementing the Mononobe-Okabe
approach (Okabe 1924, Mononobe and Matsuo 1929). For this reason, the coefficient of horizontal
acceleration (kh/g) of the backfill was set at 0.6, taking into account the abovementioned mean
peak horizontal acceleration computed at the ground surface of the bridge soil profile, whereas the
coefficient of vertical acceleration (kv/g) was set at 0.42kh.
The above calculations are summarized in Table 1, where the Q and d values refer to a
triangular soil pressure distribution. The ratio (qmax/qAE) of the maximum allowable pressure (qmax)
over the active seismic earth pressure (qAE) yields a safety factor against transverse structural
stability equal to 0.7, indicating a critical transverse condition of the bridge that should be
encountered in the design of the rehabilitation measures.
7. Safety factor against bearing capacity of the foundation soil
Within a second, more rigorous analysis stage of the bridge seismic response, a three-
dimensional finite element model was analyzed in the frequency domain by implementing the
finite-element code ANSYS (ANSYS 2001). In order to reduce computational cost, a
representative part of the bridge (Fig. 10) was modelled with proper boundary conditions to obtain
equivalent modal characteristics with the whole bridge (Rovithis et al. 2006, Rovithis and Pitilakis
2011). Cubic elements were employed to reproduce the actual geometry of the bridge by assuming
perfect bonding between the elements in the framework of a stress-based analysis. The mean value
of the elastic modulus of elasticity (Es) obtained from laboratory tests at 10 GPa was adopted as a
modelling approximation of the cracked and heterogeneous material of the bridge (OPCM 3274,
2005). Both fixed- and flexible-base models were analyzed to investigate soil-structure interaction
Fig. 10 Three-dimensional finite-element model of a representative part of the “De Bosset” bridge adopte
d
for the stress-based analysis of the bridge seismic response
655
Emmanouil N. Rovithis and Kyriazis D. Pitilakis
Table 2 Modal characteristics of the “De Bosset” bridge corresponding to (a) the fixed-base model (b) the
flexible-base model before the rehabilitation measures and (c) flexible-base model after the rehabilitation
measures. Results refer to mean elastic material properties (Es=10 GPa) derived from laboratory tests
(Papayianni et al. 2007)
Mode
(a) (b) (c)
Fixed-base model Flexible-base model Flexible-base rehabilitated
model
Frequency (Hz) Mass fraction Frequency (Hz) Mass
fraction Frequency (Hz) Mass fraction
1 35.67 0.645 1.68 0.705 2.15 0.881
2 39.38 0.646 3.78 0.745 4.51 0.882
3 47.62 0.813 3.91 0.997 6.23 0.996
Table 3 Modal characteristics of the “De Bosset” bridge corresponding to (a) the fixed-base model (b) the
flexible-base model before the rehabilitation measures and (c) flexible-base model after the rehabilitation
measures. Results refer to a reduced modulus of elasticity E’s=0.1Es
Mode
(a) (b) (c)
Fixed-base model Flexible-base model Flexible-base rehabilitated
model
Frequency (Hz) Mass fraction Frequency (Hz) Mass
fraction Frequency (Hz) Mass fraction
1 11.27 0.634 1.62 0.743 2.058 0.867
2 12.45 0.635 2.46 0.75 2.896 0.868
3 15.06 0.799 3.67 0.93 5.249 0.991
effects by means of Winkler spring supports introduced at the base of the bridge piers. For this
reason, pertinent analytical formulas reported in Gazetas (1991) for computing the stiffness of
surface foundations were adopted. It is reiterated that the foundation area of the bridge piers has
plan dimensions 10 m×5 m.
7.1 Modal analysis
Modal characteristics of fixed- and flexible-base bridge models are summarized in Table 2,
referring to the first three vibrational modes with a dominant translational component in the
transverse direction. Note the particularly stiff first-mode response with a natural period at 0.03
sec under fixed-base conditions. The comparison of natural modes between fixed- and flexible-
base models reveals a strong effect of soil-structure interaction by increasing the fundamental
natural period of the bridge up to 0.6 sec for the flexible-base case. Similar observations on the
vibrational characteristics of old stone bridges have been reported in Rota et al. (2005) and
Brencich and Colla (2002) based on field measurements. The modal response of the rehabilitated
bridge discussed in the ensuing, is also reported in Table 2. Given the highly deteriorated and
heterogeneous load-resisting system of the “De Bosset” bridge, the modal response of the structure
was re-examined under the assumption of a substantially lower modulus of elasticity E’s of the
bridge equal to 0.1Es. Modal analysis results obtained for E’s are summarized in Table 3, denoting
656
Seismic assessment and retrofitting measures of a historic stone masonry bridge
lower values of eigenfrequencies. However, strong effects of soil-structure interaction are
observed in both cases, whereas comparable the first-mode natural frequencies are obtained under
flexible-base conditions. For this reason, the seismic response of the bridge reported in the
following sections corresponds to the modal response reported in Table 2, where Es is set at 10
GPa.
7.2 Response spectrum analysis
Having identified the vibrational characteristics of the bridge, a series of response spectrum
analyses were performed to estimate the vertical stress field under the bridge piers. The mean
acceleration response spectra specified at the base of the bridge models were those computed from
the free-field ground response analyses. Code-defined elastic spectra were also implemented as a
conservative loading in the transverse direction (Rota et al. 2005). The mean vertical stress
developed under the foundation area of a typical pier was compared with the bearing capacity of
the foundation soil which in turn was computed from the available geotechnical data. The
corresponding safety factor (SFsoil), referring to the ratio of the bearing capacity of the bridge
foundation soil over the mean vertical stress developed under the bridge piers, is plotted in Fig. 11,
for fixed- and flexible-base conditions, respectively. Each bar in the graph refers to a different
loading spectrum obtained from soil response analysis (“A1-A4”, “A5-A9” and “A10-A14” soil
profiles) and code regulations (EAK2000 elastic spectrum for soil category D and EC8 elastic
spectrum for soil category C). The deviation between fixed- and flexible-base model reflects the
effect of soil-structure interaction on the vibrational characteristics of the structure and
consequently on the imposed seismic loading. With reference to the pre-rehabilitated bridge, the
safety factor related to the bearing capacity of the bridge foundation soil is below unity for both
fixed- and flexible-base cases.
8. Rehabilitation measures
The computed safety factors against transverse structural strength and bearing capacity of the
foundation soil indicated clearly that the retrofitting scheme should be oriented towards structural
strengthening and foundation soil improvement of the bridge. Upon summarizing observations
from the field inspections, the comprehensive field and laboratory tests and the seismic response
analysis of the pre-rehabilitated bridge, it was decided that the retrofitting scheme should take into
account:
• The restoration of the original masonry authentic nature that was substantially corrupted by
the interventions with reinforced concrete after the strong M7.2 earthquake of 1953.
• The extensive longitudinal and transversal cracks in the tensile zones under the arches.
• The deep corrosion, loss of mass and poor condition of joint mortars in the authentic stone
material.
• The delamination of the concrete sections and detachment of the covering.
• The scoured foundation of the bridge piers.
• The inadequate transverse strength of the bridge in case of a strong earthquake.
• The low-strength and the insufficient bearing capacity of the compressible foundation soil.
In this regard, the adopted rehabilitation scheme shown in Fig. 12 was based on the following
design concepts: (a) maintenance of the monumental features of the bridge by partial or total
657
Emmanouil N. Rovithis and Kyriazis D. Pitilakis
restoration of the detached stones and the concrete facades with local stones being compatible with
the authentic material, (b) increase of the bridge transverse strength with highly-resistant mortar to
connect the new stone elements and closely spaced stainless steel lateral tendons to confine the
bridge spandrel walls along the longitudinal direction, (c) improvement of the foundation soil by a
group of micropiles, designed as a complementary load-transfer mechanism, while increasing the
bearing capacity of the foundation soil and (d) protection of the piers base against future scouring
with stone-gravel material.
With reference to the foundation soil improvement, the final design of the rehabilitation
measures (Pitilakis 2006) yielded for each bridge pier a group of twenty-two micropiles, drilled
from the bridge deck until the stiff gravel-sand layer found at -10 m (Fig. 6). The pile diameter
(D), the normalized pile spacing (S/D) and the pile length to pile diameter ratio (L/D) were set at
0.25 m, 8 and 60, respectively. The axial load capacity of each pile is 340 KN and 270 KN,
referring to pile shaft friction and pile tip resistance, respectively, whereas the design axial load for
each pile was computed at 300 KN for combined gravity and seismic loads. Given that shaft
friction is considered as the main resisting mechanism of the micropiles, within the context of the
proposed soil strengthening intervention, the above pile group configuration provides safe
transmission of the bridge loads, by increasing the overall bearing capacity of the soil-micropiles
group system. Piles longitudinal reinforcement detailing is composed of 12 steel bars of 16 mm
diameter whereas piles transverse reinforcement is formed by 8 mm diameter spiral stirrups placed
every 10 cm along the pile shaft. More densely-spaced stirrups were designed for the piles
transverse reinforcement at the level of the bridge base and the interface between soil layers with
different stiffness. Regarding strengthening of the bridge monolithic behaviour in the transverse
direction, the design of the rehabilitation measures resulted in a series of 20 mm diameter stainless
0
0.5
1
1.5
2A1-A4
A5-A9
A10-A14
EAK2000 (Soil D)
EC8 (Soil C)
Fixed-base
model Flexible-base
model
SF
soil
Flexible-base
rehabilitated model
Fig. 11 Safety factor against bearing capacity of the bridge foundation soil; fixed-base model, flexible-
b
ase
bridge model before the rehabilitation measures and flexible-base model after the rehabilitation measures.
Each bar refers to a different spectrum loading derived from soil response analysis (A1-A4, A5-A9 and A10-
A14 soil profiles) and code regulations (EAK2000 elastic spectrum for soil category D and EC8 elastic
spectrum for soil category C)
658
Seismic assessment and retrofitting measures of a historic stone masonry bridge
2.00-3.00m
New RC slab Stainless steel bolts
RC micropiles
(L=15m, D=0.25m)
Lateral tendons
Fig. 12 Typical cross-section of the rehabilitated “De Bosset” bridge showing the proposed intervention
measures
steel lateral tendons placed every 1.5 m along the longitudinal axis of the bridge and a new lightly-
reinforced concrete slab of 15 cm thickness connected to the original bridge structure through a
mesh of stainless steel bolts, allowing for light traffic on the bridge deck. The connection between
the new reinforced concrete slab and the piles was deliberately avoided to ensure low bending
moments at the piles heads.
The proposed intervention scheme was approved by the Directorate for the Restoration of
Byzantine and Post-byzantine Monuments of the Greek Ministry of Culture and was recently
659
Emmanouil N. Rovithis and Kyriazis D. Pitilakis
applied at the bridge site. During implementation of the rehabilitation measures, a number of
unfavourable site conditions were revealed. More specifically, it was found that the filling material
between the longitudinal spandrel walls was extremely heterogeneous composed of soil, sparse
steel bars and debris (Fig. 13(a)), making the drilling and the construction of the micropiles
particularly difficult. In order to stabilise the sidewalls of the boreholes, steel tubes of 30 cm
diameter were employed to prevent debris fall inside the borehole (Fig. 13(b)). Piles reinforcement
was introduced through PVC tubes placed within the steel tubes.
The PVC tubes were not removed after casting the micropiles to protect them from strong sea
currents. The above technical solution allowed drilling of the micropiles group based on the
original design (Fig. 14(a)). For the old stone sections of the bridge, the stainless steel lateral
tendons were anchored to new stone walls constructed in the inner side of the stone facades (Fig.
14(b)). With reference to the architectural interventions, the original stone facades of the bridge
Fig. 13 (a) Filling material of the bridge composed of soil, sparse steel bars and debris that was revealed
during restoration works (b) Steel tubes embedded consecutively to prevent debris fall inside the borehole
during the drilling of the micropiles
Fig. 14 (a) Group of micropiles drilled from the bridge deck (b) Steel plate embedded in the lateral façades
of the bridge for anchoring the stainless steel tendons
660
Seismic assessment and retrofitting measures of a historic stone masonry bridge
were conserved whereas the concrete facades were repaired with corrosion -resistant mortars and
covered with a thin stone layer of 6 cm thickness.
8.1 Safety factor against bearing capacity of the foundation soil for the rehabilitated bridge
Focusing on the bearing capacity of the bridge foundation soil, the finite-element model of the
bridge was reanalyzed taking into account the group of micropiles. Upon assuming negligible pile-
to-pile interaction due to the large normalized pile spacing, the micropiles were modelled by a set
of point spring supports computed from available analytical solutions for single-pile stiffness
(Pender 1993). Modal characteristics of the rehabilitated bridge model are summarized in the
column (c) of Table 2. Naturally, the incorporation of the micropiles group increased the overall
stiffness of the soil-bridge system, leading to higher effective natural frequencies with respect to
the flexible-base case before the interventions. A new series of response spectrum analyses of the
soil-micropiles-bridge model was performed to compute the mean vertical stress developed at the
foundation of the bridge piers under the same loading scenarios. The results were compared to the
bearing capacity of the soil-micropiles system leading to the corresponding safety factors plotted
in Fig. 11 for the rehabilitated flexible-base model. The favourable effect of the combined
intervention scheme is reflected in the increase of the safety factor against bearing capacity of the
foundation soil above unity with respect to the pre-rehabilitation phase.
9. Performance of the rehabilitated bridge under the strong Cephalonia earthquakes of
01/26/2014 and 02/03/2014
On January 26th (13:55GMT) and February 3rd (03:08GMT), 2014 two strong earthquakes of
magnitude M6.1 and M6.0, respectively, stroke the Island of Cephalonia, inducing major
geotechnical failures and damages on structures and lifelines. Particularly high ground
accelerations (up to 0.70 g) were recorded at the peninsula of Paliki (GEER/EERI/ATC Report
2014, Theodulidis et al. 2016) which is less than 15 km away from the “De Bosset” bridge, by
permanent accelerometric stations and a temporal accelerometric network deployed after the first
earthquake of 01/26/2014 by the EPPO-ITSAK team (EPPO-ITSAK reports 2014a, 2014b).
Acceleration time histories of the ground motion recorded at the Argostoli (ARG2) station are
shown in Figs. 15(a) and 15(b) for the main earthquake events of January 26th and February 3rd,
2014, having peak ground accelerations at 0.35 g and 0.26 g, respectively. The above main events
were followed by a large number of aftershocks with maximum magnitude in the order of M5.4.
The rehabilitated bridge was inspected soon after the main earthquake events (Rovithis et al.
2014). The inspection revealed no damage or visible defects at the bridge structure (Fig. 16)
indicating an overall satisfactory seismic performance, despite the large ground accelerations
experienced in this site. On the contrary, the quay wall adjacent to the “De Bosset” bridge founded
on the same soil conditions suffered permanent horizontal displacement of 10cm towards the
shoreline with an approximate backfill settlement of 15 cm after the first earthquake event of
01/26/2014. The observed lateral movement of the quay wall and the settlement of the backfill
were further increased, almost doubled at some locations, after the second earthquake of
02/03/2014 (Fig. 17).
Given the comparable geological setting between ARG2 and CH accelerometric station sites,
the recorded motions at ARG2 station were deconvoluted to the bedrock level by implementing
661
Emmanouil N. Rovithis and Kyriazis D. Pitilakis
-0.4
-0.2
0
0.2
0.4
0 5 10 15 20 25
t (sec)
Acc.
(g)
N-S recording_01/26/2014
(peak -0.349g)
(a)
-0.4
-0.2
0
0.2
0.4
0 5 10 15 20 25
t (sec)
Acc.
(g)
N-S recording_02/03/2014 (peak -0.264g)
(b)
Fig. 15 Acceleration time histories (N-S component) recorded at Argostoli (ARG2) station for (a) the
mainshock of 01/26/2014 (13:55GMT) and (b) the mainshock of 02/03/2014 (03:08GMT) (EPPO-ITSA
K
reports 2014a, 2014b)
Fig. 16 In-situ inspection of the rehabilitated “De Bosset” bridge after the two major Cephalonia earthquakes
of 01/26/2014 and 02/03/2014, revealing no damage or visible defects at the bridge structure
the soil profile at the CH station. Acceleration time histories computed at the bedrock level are
plotted in Figs. 18(a) and 18(b), having peak values at 0.174 g and 0.146 g for the earthquake
events of 01/26/2014 and 02/03/2014, respectively. The deconvoluted rock mothins were then
specified at the base of the bridge soil profile to compute the seismic response at the ground
surface of the bridge site by means of equivalent linear soil response analysis (Modaressi and
Foerster 2000). Figs. 18(c) and 18(d) show the computed acceleration time histories at the ground
surface of the bridge site with peak values equal to 0.371 g and 0.251 g for the first and the second
earthquake, respectively. The corresponding surface-to-bedrock peak ground acceleration ratio is
close to two, in agreement with the mean amplification ratio obtained during the design of the
rehabilitation measures (Fig. 7(b)).
In complements to the visual evidence on the effectiveness of the rehabilitation measures, the
bridge response under the seismic loading imposed by each earthquake was further explored by
662
Seismic assessment and retrofitting measures of a historic stone masonry bridge
Fig. 17 Recorded failures adjacent to the “De Bosset” bridge; permanent lateral displacement of
the quay wall and settlement of the backfill induced by (a) the first M6.1 earthquake event of
26/01/2014 (13:55GMT) and (b) the second M6.0 earthquake event of 03/02/2014(03:08GMT)
-0.4
-0.2
0
0.2
0.4
0 5 10 15 20 25
t (sec)
Acc. (g)
Computed
_
bedrock
_
01/26/2014
(peak 0.174
g
)
(a)
-0.4
-0.2
0
0.2
0.4
0 5 10 15 20 25
t (sec)
Acc. (g)
Computed_bedrock_02/03/2014
(peak 0.146g)
(b)
-0.4
-0.2
0
0.2
0.4
0 5 10 15 20 25
t (sec)
Acc. (g)
Computed_bridge soil surface_01/26/2014
(peak 0.371g)
(c)
-0.4
-0.2
0
0.2
0.4
0 5 10 15 20 25
t (sec)
Acc. (g)
Computed_bridge soil surface_02/03/2014
(peak 0.251g)
(d)
Fig. 18 (a, b) Acceleration time histories computed at the bedrock of the CH station soil profile for the two
earthquake events of (a) 01/26/2014 and (b) 02/03/2014. (c, d) Acceleration time histories computed at the
surface of the bridge soil profile for the two earthquake events of (c) 01/26/2014 and (d) 02/03/2014
663
Emmanouil N. Rovithis and Kyriazis D. Pitilakis
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
0 0.25 0.5 0.75 1 1.25 1.5
Computed_bridge site
01/26/2014 earthq.
Computed_bridge site
02/03/2014 earthq.
T (sec)
Sa (g
)
Tssi before the interventions
Tssi after the interventions
Fig 19. Acceleration response spectra of the two earthquake motions computed at the
foundation level of the bridge; the effective natural period (Tssi) of the flexible-base bridge
model before and after the intervention measures is also shown
0
0.5
1
1.5
2
2.5 Cepha lonia earthq.1
(01/26/2014)
Cephalonia earthq.2
(03/02/2014)
Flexible-base
model Flexible-base
rehabilitated model
SF
soil
Fig. 20 Effect of rehabilitation measures on the bridge safety factor against bearing capacity of the
foundation soil under the computed seismic loading induced by the two earthquakes of 01/26/2014 an
d
02/03/2014
implementing the 3D numerical model of the bridge. The corresponding acceleration response
spectra specified at the base of the bridge model are compared in Fig. 19. The effective natural
periods (Tssi) of the pre- and post-rehabilitated bridge model under flexible-base conditions are
also noted in the graph, referring to the fundamental vibrational mode (Table 2). It is observed that
664
Seismic assessment and retrofitting measures of a historic stone masonry bridge
the rehabilitated bridge sustained larger seismic loading under the second earthquake of
03/02/2014, due to the decrease of the effective natural period after the intervention measures. The
above is reflected in the safety factor against bearing capacity of the foundation soil (SFsoil)
computed for the second earthquake (Fig. 20), which is lower than the SFsoil obtained for the first
earthquake. However, for both earthquake events, SFsoil is increased above unity with respect to
the pre-rehabilitation model response, indicating the capacity of the bridge foundation system after
the interventions to withstand safely the imposed seismic loading.
10. Conclusions
A rather complicated and efficient retrofitting scheme was presented for the old stone masonry
“De Bosset” bridge founded on unfavourable soil conditions in the particularly high-seismicity
region of Cephalonia Island in Greece. The comprehensive survey and reconnaissance campaigns
and studies undertaken to evaluate the pathology and the seismic performance of the bridge were
described. Simplified methods of transverse failure assessment and stress-based finite-element
analyses were implemented to investigate the structural stability of the bridge, leading to a set of
rehabilitation measures by combining structural strengthening and foundation soil improvement.
Special attention was paid to the preservation of the architectural, archaeological and aesthetic
features of the monument. The detailed study of the bridge pathology and the design of the
intervention measures revealed the following important points:
• In-situ inspection and testing of structural materials and foundation soil for old stone masonry
bridges was particularly critical for the “De Bosset” bridge, revealing a highly deteriorated
structure founded on a low-strength and deformable soil. The above delineated the design of the
rehabilitation scheme, requiring structural strengthening and upgrade of the foundation bearing
capacity.
• Soil-structure interaction may possess a key role in controlling the vibrational characteristics
of such type of structures that should be considered during the design of the intervention measures.
In the case study of the “De Bosset” bridge, the incorporation of the soft foundation soil modified
substantially the vibrational characteristics of the massive structure, leading to a considerable
increase of its effective natural period with respect to the fixed-base case.
• The inadequate transverse strength associated with the out-of-plane failure mechanism of the
bridge spandrel walls was identified as a major detrimental factor during the assessment of the
bridge seismic stability.
• The proposed rehabilitation scheme combines foundation soil improvement with micropiles
and structural strengthening mainly with transverse stainless steel tendons connecting the spandrel
walls of the bridge. The latter contributed to the monolithic behaviour of the structure in the
transverse direction by partially disallowing out-of-plane deformation. The retrofitting measures
are complemented by the reconstruction of the masonry walls and the protection of the piers
foundation against future scouring with stone-gravel material.
• The rehabilitated “De Bosset” bridge sustained successfully the two strong M6.1 and M6.0
earthquakes that stroke the Island of Cephalonia on 26/01/2014 and 03/02/2014, respectively, with
no damages or visible defects, indicating a very satisfactory seismic performance, despite the high
ground accelerations experienced in this site. Further evidence on the efficiency of the intervention
measures was provided by numerical analysis of the bridge seismic response using ground motion
records of the two seismic events.
665
Emmanouil N. Rovithis and Kyriazis D. Pitilakis
Acknowledgements
The present study was partially funded by the research project “Strengthening and restoration
of the “De Bosset” bridge in Argostoli, Cephalonia” under the auspices of Directorate for the
Restoration of Byzantine and Post-byzantine Monuments, Greek Ministry of Culture. The Authors
wish to thank Mr. T. Vlachoulis (Director), Mrs Ioanna Karani, Mrs Efi Chorafa and Mrs Eleni
Zarogiani from the Directorate for the Restoration of Byzantine and Post-byzantine Monuments of
the Greek Ministry of Culture for providing notes and photos during restoration works and Mr.
Aris Poziopoulos who originally conceived the need for retrofitting the “De Bosset” bridge and
initiated with great courage the whole project within the Greek Ministry of Culture and the local
community. The contribution of Professors I. Papagianni and Th. Tika of AUTH is acknowledged
for performing laboratory tests and providing structural and soil strength properties of the bridge.
The Authors wish to thank also Dr. P. Apostolidis for conducting the in-situ geophysical tests and
Assistant Professor A. Sextos of AUTH for his contribution on the numerical analysis of the
bridge seismic response.
References
ANSYS Inc. (2001), ANSYS User’s manual, SAS IP, Houston, USA.
Asteris, P.G., Chronopoulos, M.P., Chrysostomou, C.Z., Varum, H., Plevris, V., Kyriakides, N. and Silva,
V. (2014), “Seismic vulnerability assessment of historical masonry structural systems”, Eng. Struct., 62,
118-134.
AUTH-ITSAK (1996), Final report on Study of the effect of local soil conditions, geomorphology and soil
structure interaction on the recordings of the National Accelerographic Network.
Barros, F. and Luco, J. (1995), “Identification of foundation impedance functions and soil properties from
forced vibration tests on Hualien containment model”, Soil Dyn. Earthq. Eng., 14(4), 229-248.
Brencich, A. and Colla, C. (2002), “The influence of construction technology on the mechanics of masonry
railway bridges”, Proceedings of the 5th International Conference on Railway Engineering, London.
Cakir, F. and Seker, B. (2015), “Structural performance of renovated masonry low bridge in Amasya,
Turkey”, Earthq. Struct., 8(6), 1387-1406.
CEN (2002), European Committee for Standardization Eurocode 8 Design of Structures for Earthquake
Resistance, Part 5: Foundations, Retaining structures and Geotechnical Aspects, Brussels.
Code of Practice for Structural Use of Masonry (1992), Part 1: Unreinforced Masonry, BS 5628-1, British
Standards Institution, London,.
ΕΑΚ2000, Greek Seismic Code, OASP, Athens.
EPPO-ITSAK (2014a), “The earthquake of 26/01/2014 (M6.1) in Cephalonia (Greece): Strong ground
motion, soil behaviour and response of structures”, Available online at
http://www.itsak.gr/en/news/news/70.
EPPO-ITSAK (2014b), “Strong ground motion of the February 3, 2014 (M6.0) Cephalonia earthquake:
Effects on soil and built environment in combination with the January 26, 2014 (M6.1) event”, Available
online at http://www.itsak.gr/en/news/news/79.
Erdogmus, E. and Boothby, T. (2004), “Strength of spandrel walls in masonry arch bridges”, Transport. Res.
Record: J. Transport. Res. Board, 1892, 47-55.
Gazetas, G. (1991), “Formulas and charts for impedance functions of surface and embedded foundation”, J.
Geotech. Eng., ASCE, 117(9), 1363-1381.
GEER/EERI/ATC (2014), “Earthquake Reconnaissance January 26th/February 3rd 2014 Cephalonia”, Greece
Events, Report Version 1.
Griffith, M., Magenes, G., Melis, G. and Picchi, L. (2003), “Evaluation of out-of-plane stability of
666
Seismic assessment and retrofitting measures of a historic stone masonry bridge
unreinforced masonry walls subjected to seismic excitation”, J. Earthq. Eng., 7(1), 141-169.
Krstevska, L.S., Kustura, M. and Tashkov, L.A. (2008), “Experimental in-situ testing of reconstructed old
bridge in Mostar”, Proceedings of the 14th World Conference on Earthquake Engineering, Beijing.
Modaressi, H. and Foerster, E. (2000), Cyberquake: User’s Manual, BRGM, France.
Mononobe, N. and Matsuo, O. (1929), “On the determination of earth pressure during earthquakes”,
Proceedings of the World Engineering Congress, Tokyo.
Okabe, S. (1924), “General theory on earth pressure and seismic stability of retaining walls and dams”, J.
JSCE, 10(6), 1277-1323.
OPCM 3274. (2005), Ordinanza del Presidente del Consiglio dei Ministri n. 3274 del 20 Marzo 2003: Primi
elementi in materia di criteri generali per la classificazione sismica del territorio nazionale e di normative
tecniche per le costruzioni in zona sismica (ANNEX 11.D).
Page, J. (1993), “Repair and strengthening of masonry arches”, Proceedings of the IABSE Symposium on
Structural Preservation of the Architectural Heritage, Rome.
Papayianni, I., Pachta, V. and Alexiou, A. (2007), “Study of the constructing materials, techniques and
pathology symptoms of the stone bridge De Bosset in Kefalonia”, Proceedings of the 5th International
Conference on Arch Bridges, Madeira.
Papazachos, B. (1997), Earthquakes of Greece. Ziti publications, Thessaloniki.
Preciado, A., Bartoli, G. and Budelmann, H. (2015), “Fundamental aspects on the seismic vulnerability of
ancient masonry towers and retrofitting techniques”, Earthq. Struct., 9(2), 339-352.
Pender, M.J. (1993), “Aseismic pile foundation design analysis”, Bull. NZ. Nat’l Soc. Earthq. Eng., 26(1),
49-161.
Pitilakis, K. (2006), “Rehabilitation and Reinforcement Study of DeBosset bridge at Argostoli, Cephalonia”,
Final Report submitted to Directorate for the Restoration of Byzantine and Postbyzantine Monuments,
Greek Ministry of Culture, 240.
Proske, D. and Van Gelder, P. (2009), Safety of Historical Stone Arch Bridges, Springer Publications.
Roesset, J.M. (1976), “Soil amplification of earthquakes”, (Ed.) C.S., Desai and J.T., Christian, Numerical
methods in geotechnical engineering, New York: McGraw-Hill.
Rota, M., Pecker, A., Bolognini, D. and Pinho, R. (2005) “A methodology for seismic vulnerability of
masonry arch bridge walls”, J. Earthq. Eng., 9(2), 331-353.
Rovithis, E. and Pitilakis, K. (2011), “Seismic performance and rehabilitation of old stone bridges in
earthquake-prone areas: the case of Debosset bridge in Greece”, Proceedings of the 1st International
Conference on Bridges, Innovations on Bridges and Soil-Bridge Interaction, Athens.
Rovithis, E., Pitilakis, K. and Apostolidis, P. (2006), “Rehabilitation of the historic bridge De-Bosset in
Argostoli”, Proceedings of the 1st Hellenic Conference on restoration of monuments, ETEPAM,
Thessaloniki.
Rovithis, E., Pitilakis, K., Vlachoulis, T., Karani, I., Chorafa, E. and Zarogiani, E. (2014), “De Bosset
monumental stone bridge in Cephalonia: Strengthening measures and seismic response under the
earthquakes of 26/01/2014 and 03/02/2014”, Proceedings of the 2nd International Conference on Bridges,
Innovations on Bridges and Soil-Bridge Interaction, Athens.
Theodoulidis, N., Karakostas, Ch., Lekidis, V., Makra, K., Margaris, B., Morfidis, K., Papaioannou, Ch.,
Rovithis, Emm., Salonikios, T. and Savvaidis, A. (2016), “The Cephalonia, Greece, January 26 (M6.1)
and February 3(M6.0) earthquakes: Near field ground motion and effects on soil and structures”, Bull.
Earthq. Eng., 14(1), 1-38.
SA
667
  • ... Proposed project for structure retrofit and modification of bridge base soil has resulted in improvement of seismic behavior of Bosset Bridge. After the powerful earthquake in 2014 in Cephalonia Island, not damages have been seen in Bosset Bridge; therefore the seismic performance of the bridge was satisfying after the reinforcement and reconstruction process (Rovithis and Pitilakis 2016). Palu earthquake components which happened in 1978, was used for seismic analysis of Palu Bridge. ...
    Article
    Full-text available
    Seismic behavior of Osmanli and Senyuva stone bridges was addressed in this study. A combination of FEM and DEM was employed for getting closer to the real behavior of the bridge. One of the unique features of this combinational method is simulation close to reality. Modal numerical analysis was also used to verify the modeling. At the end of earthquake, a part of two lateral walls of Osmanli bridge was broken. The growth of arch cracks also increased during the earthquake. A part of right-hand wall of Senyuva Bridge was destructed during the earthquake. The left-hand side of the bridge wall was damaged during the earthquake but was not destructed.
  • ... The results indicated that the displacement of the masonry arch bridges increases with rise in the ratio of height to width of bridges' pier [34][35][36][37]. To obtain more information about seismic assessment of the masonry arch bridges, readers are referred to refs [38][39][40][41]. It is worthwhile to mention that arches have been used in various historical constructions besides their wide usage in masonry bridges. ...
    Article
    There is a large number of old arch bridges in Iran's railway network which the majority of them have been built more than 70 years ago. These structures, though displaying a good performance under the effect of service (gravity) loads but the seismic (lateral) loads have not been taken into account in their design; therefore, the seismic assessment of these structures is deemed necessary for obtaining their failure and seismic performance levels. So they can be strengthened or rendered out of order in case it is found necessary. There are numerous methods for evaluating the residual capacity of the structures under seismic excitation which among them the non-linear dynamic analysis or the incremental dynamic analysis (IDA) is the most accurate and perfect method. Selecting 22 far-field earthquake records, the present study investigated two plain concrete arch bridges which are located in kilometers 23 and 24 of the old Tehran-Qom railway, respectively. By using the finite element method and the macro-modeling method the km-23 and km-24 bridges were simulated and totally 231 non-linear dynamic analyses have been conducted. The results obtained from the incremental dynamic analysis indicated that the seismic behavior of the bridge with longer span length is more vulnerable (km-23) than the bridge with lower span length (km-24). Generally, the seismic performance of plain concrete arch bridges has been found inappropriate, therefore to improve their performance it is necessary to retrofit these type of structures.
  • Article
    There are many century-old plain-concrete-arch railway bridges in Iran in need of seismic assessment. Given the logistical difficulties of conducting field tests, the use of computer models is being investigated as an alternative. Three-dimensional finite-element models were created for two bridges that had already been physically tested. After calibration with the test results, the modelled bridges were then subjected to a modelled earthquake to evaluate their behaviour under seismic loading. The results demonstrate that seismic response depends entirely on the geometrical and material properties, and that bridge perform better in transverse direction.
  • Article
    Full-text available
    Intensity measures (IMs) are used as a link between seismic hazard and seismic demand analysis and therefore have a key role in performance-based earthquake engineering. To the best of our knowledge, no study has been carried out on the determination of suitable IMs to evaluate the seismic demand of plain concrete arch bridges. In the present study, the efficiency, sufficiency, scaling robustness and practicality of 34 potential IMs for evaluating the seismic response of two old railway plain concrete arch bridges in km-23 and km-24 of Tehran–Qom railway are investigated. The considered bridges are simulated using finite-element method and subjected to incremental dynamic analysis (IDA) using 22 far-field earthquake ground motion records. Complete response of the models is obtained through IDA method in terms of engineering demand parameter measured by the maximum displacement of the bridges. The optimal IMs among the considered intensity measures for evaluating seismic demand of the investigated plain concrete arch bridges are recognized using the concepts of efficiency, sufficiency, scaling robustness and practicality. Using the results of the regression analysis, it is concluded that root mean square acceleration is the optimal IM based on efficiency, sufficiency, scaling robustness and practicality for seismic response evaluation of plain concrete arch bridges under far-field ground motions.
  • Conference Paper
    Full-text available
    The stone Bridge DeBosset in Argostoli the capital of the island of Kefalonia, Greece, was firstly constructed in 1810, in order to connect above sea the two opposite coasts of the Argostoli gulf (Argostoli – Drapano). It was considered to be one of the biggest above sea stone bridges of the Mediterranean basin at those times, with a total length of 750m. It is still in service, but serious problems have been appeared. The continuous contact with the sea water and waves, the different types of materials used for interventions, as well as the increased car circulation, have resulted in severe pathology symptoms to the materials and structure. Based on the analysis of the existing building materials, new repair materials were proposed so as to be compatible with the old ones, in terms of aesthetic harmonization and functionality in the structural system.
  • Article
    Full-text available
    Ancient masonry towers constitute a relevant part of the cultural heritage of humanity. Their earthquake protection is a topic of great concern among researchers due to the strong damage suffered by these brittle and massive structures through the history. The identification of the seismic behavior and failure of towers under seismic loading is complex. This strongly depends on many factors such as soil characteristics, geometry, mechanical properties of masonry and heavy mass, as well as the earthquake frequency content. A deep understanding of these aspects is the key for the correct seismic vulnerability evaluation of towers and to design the most suitable retrofitting measure. Recent tendencies on the seismic retrofitting of historical structures by means of prestressing are related to the use of smart materials. The most famous cases of application of prestressing in towers were discussed. Compared to horizontal prestressing, vertical post-tensioning is aimed at improving the seismic behavior of towers by reducing damage with the application of an overall distribution of compressive stresses at key locations.
  • Article
    Full-text available
    The M6.1 and M6.0 Cephalonia (Greece) earthquakes on 26 January and 3 February 2014, were right lateral strike-slip events. Both shocks occurred on the Cephalonia Transform Fault zone. Strong ground motion was recorded in the near-fault at the permanent and temporary accelerograph network of ITSAK, with the highest to date acquired peak ground acceleration in Greece (PGA = 0.77 g at Chavriata-CHV1 station). Local site effects in combination with source effects, have strongly affected near-fault ground motion. Landslides, rock sliding effects, behavior of stone masonry retaining walls, road embankments, road network failures, ports and liquefaction are also investigated and presented. Seismic response of different type of structures at the stricken area is presented and comparison of the near-fault recorded ground motion with seismic code provisions in Greece is attempted. Although, near-fault seismic excitation imposed to Cephalonia buildings was much higher than the design values foreseen by the old and recent codes, corresponding damage was much lower than one could expect. The over-strength of structures together with a long established good construction practice on the island of Cephalonia could explain their favourable response to high seismic actions, overwhelming those of seismic code provisions. However, buildings constructed according to the 1959 Greek Seismic Code or earlier, should be investigated in more detail and if high vulnerability is detected, it is necessary to strengthen them according to modern seismic code provisions.
  • Article
    Methods of assessing, for preliminary design purposes, the stiffness and capacity of pile foundations under seismic forces are presented. Although the main thrust of the paper is to a seismic design the methods are applicable to other forms of dynamic excitation of pile foundations. Emphasis is placed on expressions for pile stiffness and capacity in the form of simple formulae that can be incorporated into spreadsheet or similar types of software. The use of the equations is illustrated with a number of worked examples. Where possible the methods are justified by data from field testing of foundations at prototype scale.
  • Book
    Full-text available
    Historical stone arch bridges are still a major part of the infrastructure in many countries. Although this type of bridge has proven to be an efficient construction type, it often poses the problem of insufficient numerical models of the load bearing behavior. Therefore the book introduces methods to adapt life loads and introduces different types of numerical models of the load resistance respectively. The book continues with the introduction of specific damages and strengthening techniques. The book particularly focuses on the probabilistic safety assessment of historical arch bridges, for which often only limited material and structural data is available.
  • Article
    Full-text available
    Masonry bridges are the vital components of transportation systems. Although these bridges were constructed centuries ago, they have served a purpose from ancient times to the present day. However, the bridges have needed local renovation and therefore have been rebuilt over different periods in many places. This study focuses on Low Bridge, which is an example of renovated masonry bridges in Turkey. It essentially assesses the structural behavior of the masonry bridge and investigates the integrity of the renovated components. For this purpose, the mechanical properties of the bridge material have been primarily evaluated with experimental tests. Then the static, modal and nonlinear time history analyses have been carried out with the use of finite element methods in order to investigate the structural behavior of the current form of the bridge.
  • Conference Paper
    Full-text available
    The historic stone masonry "De Bosset" bridge in Argostoli sustained the recent Cephalonia earthquake sequence of 26/01/2014 and 03/02/2014 mainshocks with no damages or visible defects. The above seismic performance of the bridge is analyzed in conjunction with the favorable effect of rehabilitation measures deployed prior to the two earthquakes.