Engineering Failure Analysis 57 (2015) 31–55

Contents lists available at ScienceDirect

Engineering Failure Analysis

journal homepage: www.elsevier.com/locate/efa

In-situ NDT testing procedure as an integral part of failure

analysis of historical masonry arch bridges
Otello Bergamo a,⁎, Giuseppe Campione b, Stefano Donadello a, Gaetano Russo a
a
Dept. Of Civil Engineering and Architecture, Univ. of Udine, Via delle Scienze 206, 33100 Udine, Italy
b
Dept. Of Civil, Environmental, Aerospace and Material Engineering, Univ. of Palermo, Piazza Marina 61, 90133 Palermo, Italy

a r t i c l e i n f o a b s t r a c t

Article history: A nineteenth-century masonry arch bridge was analyzed as an illustrative example to explain the
Received 30 March 2015 role of in-situ test campaigns in failure analysis and retrofit design. Test results were studied to
Received in revised form 20 June 2015 find out the advantages of each technique, with the aim of proposing an optimized in-situ testing
Accepted 11 July 2015
procedure. Standard static penetrometer, flat jack, thermographic and georadar in-situ tests were
Available online 17 July 2015
conducted. Traffic effects were analyzed by means of vibrational tests. The experimental analysis
performed to investigate damage on the bridge structures shows the degree of reliability offered
Keywords: by each technique in evaluating specific information and reproducing the global behavior of the
Masonry
structure.
Historical arch bridges
© 2015 Elsevier Ltd. All rights reserved.
Failure analysis
Diagnosis
Drawing
Finite element method

1. Introduction

Several roads in Italy and Europe are connected by masonry arch bridges. Throughout the centuries, traffic loads have increased
significantly, resulting in steady erosion of safety margins. Evaluating structural safety of historic masonry bridges is a key issue for
the maintenance of national and international infrastructural heritage. In Italy and Europe, several 18th century bridges are still op-
erating, e.g. “The Castagnara Bridge” in Padua, dating from 1848 [4], “The Stone Bridge” in Verona, “The Cernadela Bridge” Mondariz
in Northwest Spain etc. Venice counts about 480 masonry arch bridges today, but historical records tell us that their number was twice
as much in the past. Venice lagoon environment and soil characteristics are not adequate for the conservation of historical heritage. In
fact, Venice soil is generally made of clay layers with a low bearing capacity and variable position and thickness. Only some areas of the
lagoon feature overconsolidated clay, named “caranto”, which formed during a sea regression between about 6000 and 10,000 years
ago, and where the Serenissima Republic erected its buildings.
Historical experience demonstrates that masonry bridges and buildings, unlike the ones made of reinforced concrete and steel or
equipped with seismic isolation [13,16,19], were not built to withstand seismic actions. Therefore, sometimes, they collapsed, partially
or totally. Moreover, masonry aging leads to continuous erosion of the security level of a structure, with respect to both service and
ultimate limit states, as compared to its original conditions. However, structural damage is often caused by geotechnical or hydraulic
problems [15]. High pressure transmitted from abutment foundations to soil can produce large deformations and differential displace-
ment in abutment foundations. These effects represent a danger for masonry arch structures because they translate into a loss of hor-
izontal and rotational restraint at the arch base likely to evolve in hinge formations along the structure and eventually in collapse. This

⁎ Corresponding author.
E-mail address: st.eng@libero.it (O. Bergamo).

http://dx.doi.org/10.1016/j.engfailanal.2015.07.019
1350-6307/© 2015 Elsevier Ltd. All rights reserved.
32 O. Bergamo et al. / Engineering Failure Analysis 57 (2015) 31–55

Fig. 1. The collapse of “Ponte Verdura” in Sciacca (Sicily).

demonstrates that understanding the real causes of structural damage, and consequently performing in-situ tests, are essential steps
of retrofit design development [10].
The recent collapse in Sciacca (Sicily — Italy) on 4th February 2013 of the masonry arch bridge “Ponte Verdura”, dating from 1870,
is shown in Fig. 1. The collapse as “rigid body” of the pier is principally due to geotechnical and hydraulics causes.

Fig. 2. The “Paleocapa Bridge”.

O. Bergamo et al. / Engineering Failure Analysis 57 (2015) 31–55 33

Fig. 3. Venice area map, bridge location.

34 O. Bergamo et al. / Engineering Failure Analysis 57 (2015) 31–55

Fig. 4. Plan view of the bridge.

A case in point is the 43.0 m long “Paleocapa Bridge” in Venice, Italy, shown in Fig. 2. The poor bearing capacity of soil and increased
traffic loads seem to have generated widespread cracking throughout the bridge.
The structural assessment of masonry arch bridges is usually performed by means of visual investigations and a small number of
in-situ tests, including destructive and non-destructive tests; on the other hand, inspections of concrete and steel structures are car-
ried out extensively, by using several types of non-destructive tests. In this research study, georadar analysis, thermographic analysis
and vibrational tests were conducted on the masonry arch “Paleocapa” bridge. Their results are reported in this paper.

Fig. 5. North view of the bridge.

Fig. 6. South view of the bridge.

 O. Bergamo et al. / Engineering Failure Analysis 57 (2015) 31–55 35

Fig. 7. Natural stone elements of the bridge.

The aim of this paper is to explain the advantages and disadvantages of non-destructive and semi-destructive tests and to suggest
an optimized in-situ testing procedure as an integral part of retrofit design [1].

2. Case study: the historic masonry arch “Paleocapa” bridge

This bridge is one of the structures that engineer Pietro Paleocapa designed in the Venetian area in 1840, aimed at reorganizing the
Adige river system, as shown in Fig. 3. The bridge was built at different stages and was completed in the mid-19th century
(1845–1850).
In Fig. 4 is reported the plan view of the bridge.
The masonry arch bridge has a span of 15.00 m, with an arch thickness of 0.80 m at mid-span, arch rise of 2.15 m, arch central angle
of a = 64°, arch width of 6.96 m, carriageway width of 5.50 m and overall length of the structure of 43.00 m (Figs. 5 and 6).
Every element of the bridge is made of bricks and mortar, and the total weight is meaningful. The deck is composed of gravel filling
below the asphalt road structure. As shown in Fig. 7, there are some complementary natural stone elements.

Fig. 8. Cracks in the north wall surface.

36 O. Bergamo et al. / Engineering Failure Analysis 57 (2015) 31–55

3. Historical survey

A preliminary historical survey was carried out to understand the background of the bridge. The research revealed that the bridge
was designed by engineer Pietro Paleocapa in 1840 as part of the drainage design of the Adige river mouth area. The works were com-
pleted in 1874. Thus, the first information is that the foundation soil was, before drainage works, part of a marshland and its mechan-
ical properties were not be adequate, especially with respect to foundation settlement. Today, the bridge is an integral part of the
regional road n. 482 and traffic loads are far greater than the ones taken into account upon design. This represents a second important
piece of information.

4. In-situ tests and experimental analysis

In order to assess structural consistency, several non-destructive (georadar, thermograph, vibrational) and partially destructive
(flat jack) tests were conducted [15,18]. Continuous measuring of crack width and of bridge angle was performed.

4.1. Visual analysis — masonry degradation mapping

The geometric relief of the bridge was carried out with high precision instrumentation. An extensive visual survey was conducted.
All visible elements of the bridge were scanned for damage.
The bridge features several damaged spots, including two significantly deep injuries in the wing walls of the western abutment
and near the arc, as shown in Figs. 8 and 9. Wing walls were not perfectly vertical and, over the years, the injuries have widened,
in addition to masonry flaking, breaking of concrete reinforcing elements and displacement of steel cables. Foundation slippage has
caused widespread cracks in the masonry abutment and arch. With time, the structural performance of the bridge and its overall
wear resistance was deteriorated.
Collapse of the bridge might occur, resulting from seismic actions or other exceptional actions, but also simply from the bridge own
weight and dead loads. Cracks are found in various areas of the bridge, as shown in Fig. 10 [11,21].
The main damaged spots of abutments and arch are described and reported in Figs. 11, 12 and 13.

- North face of the western abutment: presence of cracks; presence of green algae, mosses, lichens (biological patina) particularly
near the river; mortar pulverization and degradation; masonry flacking and degradation; more or less deep cavities; no verticality
of wing walls.
- North face of the eastern abutment: mortar pulverization and degradation; masonry flacking and degradation; shrubs and root
system in some masonry areas; degradation of natural stone elements.
- South face of the western abutment: presence of cracks; mortar pulverization and degradation; masonry flacking and degradation;
shrubs and root system in some masonry areas; degradation of natural stone elements.
- South face of the eastern abutment: mortar pulverization and degradation, masonry flacking and degradation up to the wall top;
masonry areas with replaced or integrated mortar layers; degradation of natural stone elements.
- North face of the arch: presence of vertical cracks on the western side of the arch. On the eastern side, there are areas with mortar
pulverization and degradation, masonry flacking and degradation and presence of shrubs and root system.

Fig. 9. Cracks in the north surface of the western abutment.

O. Bergamo et al. / Engineering Failure Analysis 57 (2015) 31–55 37

Fig. 10. Main damaged spots.

Fig. 11. North view of the bridge – main damaged spots.

38 O. Bergamo et al. / Engineering Failure Analysis 57 (2015) 31–55

Fig. 12. South view of the bridge — main damaged spots.

- South face of the arch: presence of vertical cracks and of areas with mortar pulverization and degradation, masonry flacking and
degradation on both the eastern and western arch side. Masonry areas with shrubs and root system.
- Intrados of the arch: mortar pulverization and degradation, masonry flacking and degradation. Moisture in some areas of the arch
intrados.

4.2. Standard penetrometer tests

The historical and visual surveys suggest that the most important damage in the bridge may be due to settlements of the western
abutment. To investigate this possible cause, four standard penetrometer tests have been performed near the two abutments. The
diagram of the south-west test is shown in Fig. 14.
As supposed, standard penetration tests have revealed poor mechanical characteristics of the investigated soil. All the four tests
indicate the presence of a layer of compressible clay 12–13 m thick. The foundations of the abutments are ~ 10 m deep and there
are ~2 m of compressible clay under the abutments. Moreover the filling behind the walls of the abutments are composed of com-
pressible clay and it can be supposed that there was a vertical settlement due to vertical loads and/or an horizontal settlement due
to the horizontal thrust of the arch. This aspect will be clarified by the use of fem models.
Therefore, micropiles may be the best corrective solution [7,14].

Fig. 13. Damage legend.

 O. Bergamo et al. / Engineering Failure Analysis 57 (2015) 31–55 39

Fig. 14. South-west penetration standard test diagram.

4.3. Thermographic analysis

To investigate hidden damage [5,8,12], a thermographic analysis has been carried out. The measured ambient temperature was
13 °C and ambient moisture content was about 48%.
A FLIR B360 thermal camera was used: spatial resolution 1.3 mrad; accuracy +2 °C or +2% of the measured temperature; thermal
sensitivity b0.06 °C; detector FPA (focal plane array) 320 × 240 resolution pixel. The most representative thermal photos are reported
in Fig. 15.
40 O. Bergamo et al. / Engineering Failure Analysis 57 (2015) 31–55

Fig. 15. Most representative thermal photos.

 O. Bergamo et al. / Engineering Failure Analysis 57 (2015) 31–55 41

The thermographic analysis shows important results: it is a useful method to identify masonry cracks and flacking, but it can also
highlight moisture problems. Freeze/thaw cycles are very dangerous for masonry integrity. Moreover thermographic analysis can re-
veal non-homogeneous textures; in this case masonry seems homogeneous and in the arch there aren't materials discontinuity or
repairs.

4.4. Vibrational analysis

A vibrational analysis was performed to investigate the impact of traffic vibrations on the bridge masonry [20]. Vibration velocity
was measured under daily traffic actions, at three different spots of the bridge deck, and compared to the standard requirements fixed
to avoid structural damage induced by vibrations.
Vibration velocity was measured in three different spots of the bridge deck, one above the arch center and two above abutments.
The following values were assumed: 0.02 mm/s as trigger threshold and 10 s as registration length.
The instrumentation used for the analysis consisted of: a Vibration monitor ABEM Vibraloc, 24 bits, 3 channels (+1) (Fig. 16); a
geophone tern with 2.0 Hz damping frequency; Vibraloc P.C. v. 1.3 software; Sigview ver. 1.91 and UVSZA signal analysis software.
Fig. 17 illustrates, as way of example, a seismogram obtained during daily traffic and Fig. 18 reports the corresponding frequency
spectrum obtained by conventional FFT with Hanning window technique. In the diagrams, numbers 1, 2 and 3 correspond to V, L and
T, respectively.
The frequencies measured during daily traffic actions are always very similar to each other and lower than 10–12 Hz. Some peaks
of about 40 Hz or 100–120 Hz were registered.
The most important peaks are measured in the three directions and not only in the vertical one; as a consequence, under traffic
actions, the bridge apparently vibrates in all three directions. In all measurements, vibration velocities were always very small as com-
pared to the standard limits (0.4 mm/s ≪ 3.0 mm/s). Then, it can be concluded that damage in the bridge is not owed to traffic-induced
vibrations.
The vibrational analysis conducted in this research project exclude the dangerousness of traffic induced vibration for this kind of
structures that seems to be insensitive to these actions as said before.

4.5. Georadar analysis

Generally, georadar analysis is used to investigate underground utilities and objects. In this research study georadar analysis is in-
novatively used to investigate the presence of significant cracks or discontinuity in the masonry arch [6,17]. As described below the
analysis show that georadar is useful to describe the hidden parts of the structure identifying the presence of the gravel infill and
of the masonry abutments. Moreover these analysis were useful to exclude the presence of passing cracks in the arch.
The instrumentation used for radar scans consists of a Radar Systems (mod. Zond 12e 16-bit a/d conversion), equipped with an-
tennas operating at 500 and 2000 MHz. Instrumentation is composed of a transmission-reception unit (antenna tx–rx) and a signal
transduction and recording unit.
The working principle of the instrumentation, due to electromagnetic wave propagation in soil, is based on the physical phenom-
enon of energy partition at an interface between two materials with different dielectric properties (dielectric constant er and reflec-
tivity R).
In particular, at a physical interface, an e.m. wave generated at ground level and directed to the soil by an antenna is partially
reflected back to ground level (because of the electrical impedance differences of the two materials and of the e.m. wave inclination).

Fig. 16. Vibration monitor ABEM Vibraloc (L parallel to the bridge deck, T transversal to the bridge deck, and V vertical).
42 O. Bergamo et al. / Engineering Failure Analysis 57 (2015) 31–55

Fig. 17. Seismogram n. 1.

A monostatic configuration, Fig. 19b, uses only one antenna (Tx = Rx) for signal transmission and reception, while a bistatic con-
figuration uses two antennas (Tx + Rx) with variable offset. The offset selection allows, at first approximation, estimating electromag-
netic wave propagation speed, and then, once recording time is known, target depth. Said Vm the e.m. wave propagation velocity in a
m element, c the light velocity (3 · 108 m/s), er the element dielectric constant normalized to air, it results:

−1=2
V m ¼ c=ðer =e0 Þ : ð1Þ

In this research project, a bistatic configuration was used with a fixed offset (0.35 m for the 500 MHz antenna and 0.00 m for the
2000 MHz antenna).
 O. Bergamo et al. / Engineering Failure Analysis 57 (2015) 31–55 43

Fig. 18. Seismogram n. 1 frequency spectrum.

Defining the relative dieletric constant (or relative propagation velocity) is essential to calculate the depth (h) of a reflecting sur-
face, with a maximum error of ±10%. In fact:

c  tr
h ¼ pffiffiffiffi ffi ð2Þ
2 εr

where tr is the time interval between transmission and reception. Attenuation is the signal intensity reduction per unit of length
through a material. Generally speaking, the maximum depth of a georadar analysis may be said to depend from the investigated ma-
terial attenuation characteristic. High attenuation values are typical of materials characterized by high values of electrical conductivity,
like silt, clay, soluble crystalline materials, metals and saline water; low values are typical of crystalline rocks, gravel, sand and
demineralised water.
44 O. Bergamo et al. / Engineering Failure Analysis 57 (2015) 31–55

Fig. 19. Electromagnetic wave propagation in soil: a) wave emission angle and footprint; b) monostatic configuration target response.

Fig. 20. Bridge elevation and example of georadar profile superposition.

 O. Bergamo et al. / Engineering Failure Analysis 57 (2015) 31–55 45

Fig. 21. Georadar profile n. 2 above bridge deck — road side.

Materials with high attenuation values limit the depth of investigation; in plastic clays, for example, it is reduced to few centime-
ters and in metals it is practically null. Materials with high attenuation values are excellent targets, because they reflect most of the
incident radiation.
A profile (GPR section) is carried out by repeating the cycle of transmission and reception countless times, moving the antenna
along a predetermined direction; the processing software appropriately connects the received signals. Results are shown in GPR
sections.
GPR sections were made by using the space domain mode: a transmitter emits pulses depending on the distance traveled by the
operator, and this is achieved by using an encoder directly connected to the antenna.
Georadar profiles description: during tests, antennas with 500 and 2000 MHz frequency were used to investigate 3.50 m
depth from ground level. Nine radar lines were positioned above the bridge deck and above the abutments. Profile length
ranged from few meters to over 40 m. A brief description of the most significant profiles is reported below (Figs. 20, 21, 22
and 23).

• Profiles 1–5 above the bridge deck: these two longitudinal profiles obtained on the road side feature cracks in the abutments and in
the arch. The two lines show the presence of external masonry walls and of heterogeneous internal filling material.
• Profiles 2–4 above the bridge deck: these two profiles were carried out in the central part of the road and georadar tests revealed the
presence of abutments and arch, but no cracks or anomalies.
• Profiles 1–2–3–4 above the abutments: these profiles were conducted on road sides above the abutments and showed an anomaly
pointing to the presence of a crack. They also confirmed the presence of lateral masonry walls and internal filling composed of a dif-
ferent heterogeneous material, partially disturbed.

Fig. 22. Georadar profile n. 3 above bridge deck — road center.

46 O. Bergamo et al. / Engineering Failure Analysis 57 (2015) 31–55

Fig. 23. Georadar profile n. 4 above bridge deck — road side.

4.6. Single and double flat jack tests

The mechanical properties of the abutment masonry were investigated. In particular, it was necessary to define the in-situ mason-
ry stress–strain behavior and stress state [3].
Laboratory compression tests couldn't be carried out, due to the impossibility of sampling significant specimens of the abut-
ment masonry. Furthermore in-situ tests can better describe masonry behavior. A flat jack test was originally studied, by using
only one flat jack, to assess in-situ stress level of masonry, but it was later extended to include masonry stress–strain diagram by
using two flat jacks. The test is only slightly destructive. Earlier applications of this technique to some historic monuments clear-
ly indicated it has great potential and it proved to be the best solution to obtain reliable information about the main mechanical
characteristics of masonry structures (deformability, strength, state of stress). In masonry elements with regular and thin mor-
tar layers, the test is carried out by introducing a thin flat jack into a previously cut mortar layer. After completion of the test, the
flat jack can be easily removed and the mortar layer restored to its original condition. Assessing stress state is based on the stress
relaxation caused by a horizontal cut perpendicular to the wall surface; stress release is determined by a partial closing of the
cut, i.e. the distance after the cut is lower than before the cut. A thin flat jack is introduced in the cut and pressure is gradually
increased to obtain the distance measured before the cut. The displacement caused by the cut and the ones subsequently in-
duced by the flat jack are measured with removable extensometers before and after the cut and during the test. Said Pf the pres-
sure of the hydraulic system driving the displacement equal to the ones read before the cut, the equilibrium relation is:

S f ¼ K j  Ka  P f ð3Þ

where Sf is the stress value; Kj is the jack calibration constant (b1); Ka flat jack to cut area ratio (b 1).
A similar test to the one described above can also be used to determine the experimental stress–strain behavior of masonry. A sec-
ond cut is made, 500 mm away and parallel to the first one, and a second flat jack is introduced. The two flat jacks delimit a masonry
sample of appreciable size, to which a uniaxial compressive stress can be applied. Measurement bases for removable strain-gauge or
Linear Variable Differential Transformers (LVDTs) on the sample face provide information about vertical and lateral displacements.
This way, a compressive test is carried out on an undisturbed large-sized sample.
Several loading/unloading cycles can be performed at increasing stress levels in order to determine the masonry Young modulus at
various loading and unloading phases. It is interesting to compare these results to the stress level measured, in order to verify the ma-
sonry residual safety.
In this research project, one single flat jack test and four double flat jack tests were performed. The single flat jack was carried out
on the east face of the northern abutment. The four double flat-jack tests were performed as follows: the first on the west face of the
southern abutment; the second on the east face of the southern abutment; the third on the east face of the northern abutment; the
fourth on the west face of the northern abutment. The four double flat jack tests are illustrated in Fig. 24.
First of all, the masonry was cleaned, measurement bases were placed, and a first measurement set was taken. Secondly, a mortar
layer was cut by means of a diamond grinding wheel (diameter 414 mm), and another measurement set was taken. A second cut was
made parallel to the first one, and another measurement set was taken. Finally, a flat jack was introduced in each cut. Hydraulic flat
jacks of the type Boviar semi-oval MP-8A 350 × 260 × 4 mm (sheet thickness 0.8 mm) were installed and operated with a hydraulic
manual pump.
Pressure measurements were made via a pressure transducer “GETRAN”, with precision of 0.3% FSO, connected to a digital data ac-
quisition system.
 O. Bergamo et al. / Engineering Failure Analysis 57 (2015) 31–55 47

Test 1: Southern abutment, west face Test 2 : Southern abutment, east face

Test 3: Northern abutment, east face Test 4: Northern abutment. west face
Fig. 24. Double flat jack tests in the abutment walls.

Fig. 25. Stress–strain diagram — double flat jack tests.

48 O. Bergamo et al. / Engineering Failure Analysis 57 (2015) 31–55

5000
4592
4484
4500
4169
4069
4000 3751
3550

MASONRY YOUNG MODULUS (MPa)

3399
3500

3000 2758

2500

2000

1500

1000

500

0
1' 1'' 2' 2'' 3' 3'' 4' 4''

TEST AND LOAD CYCLE NUMEBER

Fig. 26. Average Young modulus values obtained from the double flat jack tests.

At different pressure stages, various measurement sets were taken as compared to the measurement bases by means of a Mitutoyo
deformometer (resolution: 1/1000 mm). A stress–strain diagram was eventually obtained. The masonry Young modulus was then cal-
culated as the average stress/strain ratio, as shown in Fig. 25.
A correction coefficient Km must be applied to pressure values. Km represents the pressure percentage value likely to be transferred
to the material surface under investigation (e.g. masonry). In simpler terms, under ideal conditions of usage, Km means the percentage
of pressure applied to a jack likely to be transferred to the material surface (masonry). Km = 0.89 was assumed for the jacks used in the
tests. A coefficient C = 0.102 N/(mm2·bar) was applied to pass from bar to N/mm2.
The stress/strain diagrams obtained from the tests are represented in Fig. 11, where the average elasticity modulus values are es-
timated. The average Young modulus values obtained from the tests are illustrated in Fig. 26.
A single flat jack test was also performed to estimate stress value at the base of the abutment wall.
First of all, three pairs of deformometer base were placed and a first set of measurement was taken by means of a digital Mitutoyo
(accuracy 1/1.000 mm). Afterward, a horizontal cut was made and a second set of measurement was taken.
Subsequently, a flat jack (type Boviar semi-oval MP-5A 350 × 260 × 4 mm, sheet thickness 8 mm) was introduced in the cut and
the necessary pressure was applied to obtain the same measurements as observed before cutting. Pressure values were recorded. The
pressure value required to obtain initial displacement, Preg = 4.5 bar, was estimated by using a “GEFRAN” pressure transducer (0.3%
FSO precision), which was connected to a data logger equipped with a monitor.

Fig. 27. Fem model of the bridge — masonry structure.

 O. Bergamo et al. / Engineering Failure Analysis 57 (2015) 31–55 49

non-structural mass

infill

Fig. 28. Fem model of the bridge — infill and non-structural mass (road asphalt structure).

Amart 
σ ¼ Amart −  C  K m  P reg ð4Þ
Asa

778:56 N N
σ¼  0:102  0:89  4:50bar ¼ 0:35 ð5Þ
908:56 mm2  bar mm2

where:

Amart: flat-jack area — 778.56 mm2

Amart*: flat-jack area in excess with respect to cut — 0 mm2
Asa: cut area — 908.56 mm2.

The elastic modulus values obtained from the tests suggest good mechanical characteristics of the investigated masonry. No tests
showed masonry collapse due to insufficient contrast.
As shown before the compression stress in the masonry of the abutments is lower than 1/10 of the stresses reached during the
double flat jack tests.
In the following section an investigation of the reliability of the result is made by finite element models.

5. Finite element models

In this section are reported some results obtained from finite element analysis useful to understand the reliability of the in situ
tests [2,9]. First of all, are presented three different models which represent three different possible boundary conditions: 1. no

masonry

node restraints
face support

Fig. 29. Fem model of the bridge — boundary conditions (node restraint and brick face support).
50 O. Bergamo et al. / Engineering Failure Analysis 57 (2015) 31–55

Fig. 30. Fem model of the bridge — position of the flat jack tests.

settlement of the abutments; 2. vertical settlement of west abutments; 3. horizontal settlement of west abutments. Subsequently is
presented the result of modal analysis and its interpretation for each cases examined.
The three models are the same except some boundary conditions ax explained below. The models are made by 36913 brick tetra
10 nodes elements, 5958 of which are used to model the gravel infill (Figs. 27–28–29). Elastic properties of the masonry modeled, in
agreement with data derived experimentally, are the follows: E = 4160 MPa, n = 0.2, g = 1800 kN/m3. Elastic properties of the gravel
infill modeled are the follow: E = 100 MPa, n = 0.35, g = 2000 kN/m3. Boundary conditions are modeled as translational restraint in
the three direction of the nodes at the base of the abutments. The effect of the earth behind the abutment walls are modeled as brick
face support with a value 10 MPa/m. The dead load of the road asphalt structure is assumed 300 daN/m2.
The position of the stress measurement in the fem models is represented in Figs. 30–31–32.

5.1. Model 1: fixed base of the abutments

Here are reported vertical displacements (Fig. 33) of the bridge due to dead loads. It can be observed that vertical stress in the po-
sition of the flat jack M3, dashed line in Fig. 34, is sensibly lower that one measured in the in-situ test (0.35 MPa).

5.2. Model 2: vertical settlement of the west abutment (imposed displacement 80 mm)

In this case a vertical displacement Dy = −80 mm is imposed to the base of the west abutment. The results show that the vertical
tension in the fem model in the position of the flat jack M3 is compatible with the in-situ measurement but the global deformation is
not compatible with the actual situation and crack pattern. In Fig. 35 are reported vertical displacements while in Fig. 36 are reported
vertical stresses in the position of the flat jack tests. Moreover in the Fig. 37 are reported principal tensile stresses: it is evident that this
stress state is not realistic because of the high value of tension in the intrados of the arch.

Fig. 31. Fem model of the bridge — west abutment — position of the stress measurement.
 O. Bergamo et al. / Engineering Failure Analysis 57 (2015) 31–55 51

Fig. 32. Fem model of the bridge — east abutment — position of the stress measurement.

5.3. Model 3: horizontal settlement of the west abutment (imposed displacement 5.5 mm)

In this case an horizontal displacement of 5.5 mm is imposed to the base of the west abutment. Vertical displacements are reported
in Fig. 38. The results show that the vertical tension in flat jack M3 (Fig. 39) is compatible with the in-situ measurement and the global
deformation is compatible with the actual situation and crack pattern. In Fig. 40 are reported principal tensile stresses.
The fem analysis shown that the most probable cause of the crack pattern in the bridge is an horizontal settlement of the west
abutment. Probably, the horizontal settlement as developed together with a vertical settlement but the first is the dominant in the
bridge stress state.

6. Investigation procedure proposal for masonry arch bridges

The aim of this research project is to suggest an optimized procedure to investigate masonry arch bridges or other similar struc-
tures with some innovative techniques. In particular this work shows that an innovative application of the georadar analysis is useful
to investigate masonry arch bridge in a non-destructive way. Moreover, vibrational analysis, flat jack tests and static penetration tests
are useful to investigate causes of damages and to calibrate fem models. Finally thermographic analysis is the most reliable technique
to find out moisture, discontinuity and non-homogeneity in the material.

Fig. 33. Fem model of the bridge — case 1: fixed base — vertical displacements.
52 O. Bergamo et al. / Engineering Failure Analysis 57 (2015) 31–55

Fig. 34. Fem model of the bridge — case 1: fixed base — vertical stress (ref. Figs. 31–32).

This case study was chosen to explain the advantages and disadvantages of each investigation technique, according to the exper-
imental results obtained during the research. A standard procedure aimed at investigating all structural aspects of a masonry arch
bridge should include the following steps: historical survey, visual survey, geotechnical survey, thermographic analysis, georadar
analysis, single and double flat jack tests and, if appropriate, vibrational and dynamic tests and analyses.
To investigate the structural health and safety of masonry arch bridges, the following techniques were applied and studied.

- A visual survey represents the first approach to estimate the structural health and safety of a bridge and to find out visible damage
and defects, but also to examine other structures in the area near the bridge and detect any possible similar damage;
- An historical survey is useful to understand the original construction techniques and materials, the geotechnical and hydraulic en-
vironment prior to bridge construction, the design approach and some other information that may be useful to describe the struc-
ture. Furthermore, an historical survey could be extended to similar structures to evaluate similar damage conditions, and their
causes and consequences;
- Geometric and topographic surveys are fundamental to model the bridge and the area;
- Single and double flat jack tests are necessary to define the stress state of the masonry, as well as to measure the average elastic
modulus required to define and verify f.e.m. models. In-situ flat jack tests are more efficient than laboratory compressive tests;
in fact, in the latter, samples are subject to altered conditions during tests and no information about real stress states can be

Fig. 35. Fem model of the bridge — case 2: vertical settlement — vertical displacements.
 O. Bergamo et al. / Engineering Failure Analysis 57 (2015) 31–55 53

Fig. 36. Fem model of the bridge — case 2: vertical settlement — vertical stress (ref. Figs. 31–32).

obtained. a georadar analysis shows great potential, because it can reveal hidden materials and damage through the whole bridge
thickness, without the need for any demolitions. Also, it offers the chance of investigating all the bridge deck and abutments;
- A thermographic analysis, unlike the georadar analysis, allows better investigating superficial damage and material discontinuity.
Therefore, in the case of a masonry arch bridge, it proves less useful than georadar analysis. However, a thermographic analysis can
reveal damage where georadar analysis cannot, e.g. lateral photo of abutments. Furthermore, a thermographic analysis is very use-
ful to find out moisture problems causing masonry flaking and degradation;
- Vibrational tests allow comparing traffic-induced vibration with the standard limits, and assessing whether traffic is the cause of
damage.

This research project demonstrates that thorough understanding of the structural health and safety of a bridge is only possible
with a complete procedure of investigation, as suggested here. The study of a wide number of masonry arch bridges highlighted
often foundational problems and the research conducted suggests adequate investigation of the soil, both superficial and deep, to
avoid engineering failures. In the case of retrofitting of existing bridges, design errors could be avoided by ample campaign of exper-
imental tests, as the one proposed here. Other works [4] show that a static load bearing test is another effective non-destructive test
for masonry arch bridges, while a dynamic characterization test is less helpful in studying massive structures. Results of tests have to
be used to analyze and understand the tension history and they have to be taken in account considering the real behavior of the

Fig. 37. Fem model of the bridge — case 2: vertical settlement — principal stress 11.
54 O. Bergamo et al. / Engineering Failure Analysis 57 (2015) 31–55

Fig. 38. Fem model of the bridge — case 3: horizontal settlement — vertical displacements.

masonry elements and structures. In this case study the stress state of the abutments was evaluated experimental by flat jack tests.
The linear analysis conducted are representative of serviceability limit state, object of the study, in fact the experimental stresses mea-
sured in the abutments coincide to the ones calculated by the fem analysis. Non-linear behavior of the masonry bridge will be inves-
tigated in the subsequent step of retrofit design where the collapse mechanisms are studied to define the correct retrofit solutions.

7. Conclusions

Several historic masonry arch bridges are still used in Italian and European roads today. The structural efficiency of this kind of
structures helps them tolerate increased traffic loads. Furthermore, masonry arch bridges are an important part of the historic and ar-
chitectural heritage. For the reasons above, it is important to preserve and not replace these structures.
Retrofit design represents the solution to rehabilitate these structures in compliance with standard requirements. It must be noted
that the first step of retrofit design is investigating the structure and understanding its structural health and safety.
In this research project, a case study was examined and the most important in-situ tests and methods of analysis were applied and
studied. Their advantages and disadvantages were explained. An investigation procedure to study masonry arch bridges is presented,

Fig. 39. Fem model of the bridge — case 3: horizontal settlement — vertical stress (ref. Figs. 31–32).
 O. Bergamo et al. / Engineering Failure Analysis 57 (2015) 31–55 55

Fig. 40. Fem model of the bridge — case 3: horizontal settlement — principal stress 11.

including the following steps: historical survey, visual survey, geometric and topographic survey, single and double flat jack tests,
georadar analysis, thermographic analysis and, if necessary, vibrational and dynamic characterization tests and static load bearing test.
In-situ tests are also useful to prove the efficiency of retrofitting works.

References

[1] A.M. Alani, A. Morteza, G. Kilic, Integrated health assessment strategy using NDT for reinforced concrete bridges, NDT&E Int. 61 (2014) 80–94.
[2] A. Bayraktar, A. Altunişik, F. Birinci, B. Sevim, T. Türker, Finite-element analysis and vibration testing of a two-span masonry arch bridge, J. Perform. Constr. Facil.
24 (1) (2010) 46–52.
[3] G. Bartoli, M. Betti, Cappella dei Principi in Firenze, Italy: experimental analyses and numerical modeling for the investigation of a local failure, J. Perform. Constr.
Facil. 27 (2013) 4–26 (SPECIAL ISSUE: Analysis of Structural Failures Using Numerical Modeling).
[4] O. Bergamo, G. Russo, S. Donadello, Retrofitting of the historic Castagnara Bridge in Padua, Italy, with fibre reinforced plastic elements, Struct. Eng. Int. 24 (4)
(November 2014) 532–543.
[5] F. Cluni, D. Costarelli, A.M. Minotti, G. Vinti, Enhancement of thermographic images as tool for structural analysis in earthquake engineering, NDT&E Int. 70
(2015) 60–72.
[6] N. Diamanti, A. Giannopoulos, M.C. Forde, Numerical modelling and experimental verification of GPR to investigate ring separation in brick masonry arch bridges,
NDT&E Int. 41 (5) (2008) 354–363.
[7] H. El-Mahmoud, E. Rufini, Gu, Micropile application for seismic retrofit of the Richmond-San Rafael toll bridge, Geotech. Eng. Transp. Proj. (2004) 1290–1298.
[8] C. Hing, U. Halabe, Nondestructive testing of GFRP bridge decks using ground penetrating radar and infrared thermography, J. Bridg. Eng. 15 (2010) 391–398
(SPECIAL ISSUE: Bridge Inspection and, Evaluation).
[9] J. Kishen, A. Ramaswamy, C. Manohar, Safety assessment of a masonry arch bridge: field testing and simulations, J. Bridg. Eng. 18 (2) (2013) 162–171.
[10] J. Lai, S. Wu, C. Chiang, Evaluating the compaction quality of backfills by stress wave velocities. Contemporary topics on testing, modeling, and case studies of
geomaterials, pavements, and tunnels, GeoHunan 2011. , American Society of Civil Engineers, 2011. 92–99, http://dx.doi.org/10.1061/47626(405)12 (ISBN
(print): 978-0-7844-7626-0).
[11] T. Li, D.P. Almond, D.A.S. Rees, Crack imaging by scanning pulsed laser spot thermography, NDT&E Int. 44 (2) (2011) 216–225.
[12] K. Maser, W. Roddis, Principles of thermography and radar for bridge deck assessment, J. Transp. Eng. 116 (5) (1990) 583–601.
[13] G. Russo, O. Bergamo, L. Damiani, Il Viadotto di Silea in Veneto: Verifica Sismica Secondo la Normativa Italiana e l'Eurocodice, Ing. Sismica 3/2008 (2008) 24–35.
[14] G. Russo, O. Bergamo, L. Damiani, Retrofitting a short span bridge with a semi — integral abutment bridge: the Treviso bridge, J. Int. Assoc. Bridg. Struct. Eng. 19
(2) (May 2009) 137–141.
[15] G. Russo, O. Bergamo, L. Damiani, D. Lugato, Experimental analysis of the “Saint Andrea” Masonry Bell Tower in Venice. A new method for the determination of
“Tower Global Young's Modulus E”, Eng. Struct. 32 (2) (2009) 353–360 (01/2010).
[16] G. Russo, O. Bergamo, S. Donadello, “Il viadotto di Dolcè”: analisi sismica delle pile secondo la normativa italiana e l'Eurocodice, Ing. Sismica XXVII/2 (2010)
49–61.
[17] M. Solla, H. Lorenzo, F.I. Rial, A. Novo, GPR evaluation of the Roman masonry arch bridge of Lugo (Spain), NDT&E Int. 44 (1) (2011) 8–12.
[18] V. Srinivas, S. Sasmal, K. Ramanjaneyulu, K. Ravisankar, Performance evaluation of a stone masonry–arch railway bridge under increased axle loads, J. Perform.
Constr. Facil. 28 (2) (2014) 363–375.
[19] P. Tsopelas, M. Constantinou, Study of elastoplastic bridge seismic isolation system, J. Struct. Eng. 123 (4) (1997) 489–498.
[20] H. Ward, Traffic generated vibrations and bridge integrity, J. Struct. Eng. 110 (10) (1984) 2487–2498.
[21] B. Weekes, D.P. Almond, P. Cawley, T. Barden, Eddy-current Induced Thermography — Probability of Detection Study of Small Fatigue Cracks in Steel, Titanium
and Nickel-based Superalloy, 2012.