You are on page 1of 8

ARTICLE IN PRESS

NDT&E International 43 (2010) 62–69

Contents lists available at ScienceDirect

NDT&E International
journal homepage: www.elsevier.com/locate/ndteint

Application of acoustic techniques in the evaluation of heterogeneous


building materials
S. Fais a, G. Casula b,
a
DIGITA—Dipartimento di GeoIngegneria e Tecnologie Ambientali, Piazza D’Armi 19, 09123 Cagliari, Italy
b
Istituto Nazionale di Geofisica e Vulcanologia, Sezione di Bologna, Via Donato Creti 12, 40128 Bologna, Italy

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

Article history: In this paper, we present the results of an application of in situ non-destructive acoustic techniques to
Received 10 December 2008 analyse the state of preservation of a colonnade of an ancient Church in the Central Sardinia (Italy). In
Received in revised form particular, low-frequency ultrasonic tests and seismic longitudinal wave transmission tomography
6 October 2009
(SLWTT) techniques were applied on the columns of the Church in order to provide information on their
Accepted 13 October 2009
state of preservation and to detect possible defects, such as zones of weakness, damages and cracks in
Available online 23 October 2009
its inner parts that cannot be visually examined.
Keywords: In order to improve the velocity input model for the SLWTT, a new experimental procedure based on
Acoustic techniques the calculation of the cross-correlation function was proposed and usefully applied.
Longitudinal wave
An estimate of the accuracy of longitudinal wave velocity measurements by SLWTT methods was
Ultrasound
carried out by application of a standard error analysis based on the law of propagation of uncertainty.
Seismic tomography
Based on the results obtained with the ultrasonic and seismic investigations, we located microcores
that confirmed information from the above non-invasive investigations.
& 2009 Elsevier Ltd. All rights reserved.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
2. Experimental section . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
3. Ultrasonic and seismic investigations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
4. Seismic tomography experiment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
5. Standard error analysis of the velocity field. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
6. Discussion of the results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
7. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68

1. Introduction Nowadays these techniques are usefully applied in diagnostics


of building materials of monumental structures [1–11], because
The propagation of acoustic waves through a stone material is they give information on their state of preservation while not
becoming more and more important in defining its dynamic causing any damage whatsoever.
properties and therefore its mechanical behaviour. Techniques This paper focuses on the application of low-frequency
that involve the propagation of acoustic waves are an efficacious (24 kHz) ultrasonic techniques and seismic longitudinal wave
tool in the characterization of rock materials in several engineer- transmission tomography (SLWTT) methods on the columns
ing applications. (Fig. 1) of the ancient (1200 AD) Church of San Bartolomeo
(Central Sardinia—Italy) to detect damaged or altered zones inside
the investigated structures.
Because of several restoration works carried out in different
 Corresponding author. Tel.: + 39 051 4151415; fax: +39 051 4151498. ages, all the columns were found to be made up of an original
E-mail addresses: sfais@unica.it (S. Fais), giuseppe.casula@bo.ingv.it, nucleus of trachite ashlars, a first lining with concrete, and an
casula@bo.ingv.it (G. Casula). outer cladding with bricks, as seen in the section in Fig. 2.

0963-8695/$ - see front matter & 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ndteint.2009.10.004
ARTICLE IN PRESS
S. Fais, G. Casula / NDT&E International 43 (2010) 62–69 63

for SLWTT, a new experimental methodology based on the


computation of the cross-correlation function between different
data sets collected for the tomographic reconstruction was
developed. The cross-correlation function between different data
sets recorded in each column was calculated in order to obtain a
cross-correlogram. The interpretation of the cross-correlogram
allowed to detect the anomaly zones of transit times of long-
itudinal wave paths on investigated sections of the columns. This
interpretation allowed a realistic starting model to reconstruct the
most reliable representation of the velocity field on investigated
column sections by the well-known algebraic reconstruction
technique ART [12,13], in order to assess the presence of damages
or alteration zones within the columns.
A standard error analysis based on the law of propagation of
uncertainty was applied to the seismic velocity data in order to
obtain reliable indications on the accuracy of the method.

2. Experimental section
Fig. 1. Scheme of the colonnade of the ancient (1200 AD) Church of San
Bartolomeo (Central Sardinia—Italy).
The main aim of the survey was to evaluate objectively the
conservation state of the San Bartolomeo colonnade from the
knowledge of the elastic state of building materials. In fact, as is
known, acoustic methods are based on the principle that the
characteristics of acoustic signals strictly depend on the elastic
state of materials. In general, alterations in the material cause the
lowering of the velocity value of longitudinal elastic waves. This
physical property can be used to map the internal structure of
investigated bodies after the application of sophisticated geophy-
sical data inversion techniques.

3. Ultrasonic and seismic investigations

In view of the nature of investigated building materials and of


the structure of the columns, which are characterised by shallow
and deep inhomogeneity, the acoustic investigation was carried
out using the integrated application of low-frequency ultrasonic
(24 kHz) and seismic techniques with the aim of detecting any
mechanical discontinuities or damaged zones through the study
of the velocity anomalies in the propagation of acoustic signals.
Ultrasonic measurements were performed using the portable
ultrasonic non-destructive digital indicating tester (PUNDIT) by
C.N.S. Electronics Ltd. (London, UK) and a portable Fluke
oscilloscope that allows the acquisition, processing, and display
of digitized wave forms.
Seismic measurements were carried out by the OYO McSeis-
170f engineering seismograph (Tokyo, Japan) with an instanta-
neous floating point (IFP) amplifier. During the data acquisition
phase, different data acquisition techniques were implemented
and several tests were carried out to select the proper acquisition
geometry, such as transmitter–receiver distance (i.e. offset) and
Fig. 2. Nature of the studied materials and structure of a column of the Church of distance between receivers.
San Bartolomeo.
Ultrasonic tests were first performed by the surface or indirect
transmission method [1–3,6–8], that is, placing the transmitter
and the receiver along the same face of the column to test the
The nature and conditions of building materials of the columns conservation state of the superficial layers of building materials,
were such that the responses from ultrasonic tests alone were not according to NORMAL—22/86 suggestions [7]. A data acquisition
sufficient to define possible alteration zones inside the columns. technique, the so called step-by-step technique that had already
As a matter of fact, the penetration depth of ultrasonic long- been tested successfully in previous experimental works
itudinal pulses is strongly limited by shallow elasto-mechanical [6,8,10,14] was used. The technique was carried out along two
conditions of building materials. parallel profiles at different levels of each face of the investigated
A better estimate of the velocity distribution inside the columns (Fig. 3) using different transmitter–receiver distances
columns was made by the SLWTT technique, which allowed a (offset) to evaluate the influence of this parameter and to check
detailed reconstruction of the velocity field on the longitudinal the elastic conditions of superficial layers of building materials.
sections inside the columns. To improve the input velocity model For the ultrasonic data acquisition, silicone snug sheets were
ARTICLE IN PRESS
64 S. Fais, G. Casula / NDT&E International 43 (2010) 62–69

Fig. 3. Data acquisition scheme carried out using the step by step surface ultrasonic
Fig. 4. Velocity map of the longitudinal ultrasonic pulses by the application of the
technique characterised by a transmitter–receiver distance of 20 cm. (first column
step by step surface ultrasonic acquisition technique (first column on the left side of
on the left side of the plan).
the plan).

chosen as coupling agents because they were assessed to be a since the seismic method operates at lower frequencies, it proved
better coupling agent than vaseline [4]. to be more suited to the investigation.
The survey direction was as perpendicular as possible to the
most probable direction of cracks, as derived from the analysis of
the column outside surface. As an example Fig. 4 shows the 4. Seismic tomography experiment
ultrasonic longitudinal velocity map obtained by processing the
ultrasonic data acquired by the above step-by-step technique in a Seismic tomography data acquisition was carried out using the
face of the first column on the left side of the plan shown in Fig. 1. standard acquisition scheme shown in Fig. 5. Sources and
The map was obtained by interpolating the velocity values receivers were positioned to obtain a good coverage of ray paths
measured along the profiles, using a 0.20 m offset. As deduced associated with longitudinal elastic waves propagating through
from tests carried out in various conditions [6], the velocity values the materials. Mechanical pulse velocity from impacts generated
must be interpreted as relative since they are also affected by the by a metal hammer of about 100 g was used as the input signal.
offset. Notwithstanding this, the longitudinal ultrasonic velocity The transmitted pulses were received at the opposite side of the
map indicates variations in the elastic conditions of shallow column by high-frequency geophones located at 15 cm space
materials (bricks and concrete). In the lower sector of the intervals (Fig. 5). The signal amplitudes were strong and therefore
investigated face, the map shows a relative high velocity zone in the first arrivals were easy to pick.
correspondence with a horizontal iron reinforcement inside the In order to obtain a realistic starting velocity model as input for
concrete. The low velocity zones in the map represent defects and the algebraic reconstruction technique (ART) [12,13], we experi-
alteration zones of shallow materials, as can be also deduced mented a methodology based on the cross-correlation function
comparing the velocity map with the macroscopic visual (CCF). The travel times of ray paths measured at constant
observation of the column. distances between seismic sources and receivers, along parallel
In this case, since we were also investigating inside columns trajectories, and characterized by homogeneous angular coeffi-
thicker than 1 m, the propagation of longitudinal ultrasonic pulses cients were organised in different data sets. In brief, the used
by the direct transmission method (transmitter and receiver on procedure is essentially made up of the following main steps:
the opposite side of the investigated structure) was made more
difficult by the presence of bricks and concrete in superficial parts K Organisation of the seismic ray transit times on data sets, each
of the columns. For this reason, in order to investigate the internal formed by the transit times measured on constant source–
parts of the columns we applied tomography with the seismic receiver distances. Each data set is characterised by a
rather than the ultrasonic method to obtain better responses from different angular coefficient expressed by: m =nnDs/H, where
deep materials. The two methods differ in the use of longitudinal m= angular coefficient, n =number of intervals, Ds= distance
elastic pulses of different frequencies and intensities. In this case between receivers or sources, H= source–receiver distance.
ARTICLE IN PRESS
S. Fais, G. Casula / NDT&E International 43 (2010) 62–69 65

K Analysis of the previously obtained data sets and choice of the K The previously described general procedure has been im-
most meaningful ones for the computation of the CCF. The data proved by considering the data set relating to the transit times
set characterised by m =0 or a more meaningful one is taken as measured for minimum offset as reference, and dividing it into
reference and is cross-correlated with other data sets. Fig. 6 suitable intervals selected according to the wavelength of
shows an example of two correlated data-sets. the transit time anomalies measured on different receivers.
K Positioning of anomalies and computation of their geometry: K Building of the cross-correlogram, which is done by contouring
in general, a reference frame is adopted with the X-axis along the sets of the previously calculated cross-correlation func-
the source or receiver profiles and the Y-axis along the source– tions arranged in successive intervals of the reference data set.
receiver distance. K Interpretation of the cross-correlogram and consequent
localisation of anomalous zones within the investigated
column section.
K Tomographic reconstruction with the ART technique, using the
cross-correlogram interpretation as input model.

Thanks to the extremely large amount of data available, with


seismic tomography the longitudinal velocity values of unda-
maged materials can be detected in a statistically significant way.
Therefore variations in longitudinal velocity values detected in
longitudinal investigated sections of the above-mentioned col-
umns can be attributed to variations in elastic conditions of
building materials.
Figs. 7 and 8 show two examples of the most representative
cases that resulted from our experimental procedure. The results
are presented as grey-scale plots.
Fig. 7 shows the interpretation of the cross-correlogram
through a longitudinal section of the first column on the left side
of the Church (left) and the reconstruction of the velocity field by
ART through the same section (right). In the case displayed on the
left side of Fig. 7, we can observe anomaly zones of transit times
detected by the cross-correlogram that correspond to low velocity
zones in the tomographic reconstruction (Fig. 7 right). Within this
section the velocity anomaly zones are defined quite clearly and
characterized by lower velocities than for the surrounding
material. In particular, in correspondence to the ashlar trachite
building material, the anomaly zones show average velocity
values 30–40% lower than those for the same undamaged
materials, which were statistically deduced from the great
quantity of processed data.
Fig. 8 shows the interpretation of the cross-correlogram (left)
and the reconstruction of the velocity field (right) through
a longitudinal section of the first column on the right side of
Fig. 5. Data acquisition scheme of the seismic tomography. the Church. In this case, the results show a better homogeneity of

Fig. 6. Example of our CCF-based procedure applied to two data sets.


ARTICLE IN PRESS
66 S. Fais, G. Casula / NDT&E International 43 (2010) 62–69

Fig. 7. Example of the cross-correlogram calculated for a vertical section through the first column on the left side of the Church (left) and the corresponding tomographic
velocity reconstruction (right). Areas of low velocity represent damages inside the column.

Fig. 8. Cross-correlogram calculated for a vertical section through the first column on the right side of the Church (left) and the corresponding tomographic velocity
reconstruction (right).
ARTICLE IN PRESS
S. Fais, G. Casula / NDT&E International 43 (2010) 62–69 67

the velocity distribution denoting a more uniform medium from b) the diagonal ray paths from transmitters to receivers, set,
the point of view of its elastic status. respectively 40 cm, below and above the transmitters them-
The described methodology requires further study. Many selves (for example: T1–R3 and T3–R1).
variables are to be evaluated during the processing. An imple-
mentation of this methodology will be performed in the future. The results of this computation are shown in Table 1, for the
Therefore the experiments carried out so far suggest that the undamaged and damaged columns. In the first four rows for cases
cross-correlation function can give an useful contribution in (a) and (b) described previously we represented the weighted
improving the input model for tomographic reconstruction. averages and standard deviations of the transit times, their
Based on the results of the experimental procedure, two percentage errors, the corresponding mean velocity and its
microcores were planned and located for data calibration. standard deviation computed with the help of Eq. (2).
The result of the same computation is shown for the damaged
column in the last four rows of the same table. Moreover, taking
5. Standard error analysis of the velocity field into account these results we can conclude that: as is known, the
undamaged column is characterised by higher mean velocities for
In order to estimate the accuracy of the method used to direct paths (2816+ /  173) m/s, compared to the values com-
compute the velocity field of longitudinal sections of the columns, puted for the damaged one (2679+ /  166) m/s.
we applied a standard error analysis; in other words we applied The main part of the root mean square (RMS) error of the
the law of propagation of uncertainty [16] to the equation: v= x/t, velocity field is represented by the term that contains the RMS of
expressing mean velocity as a ratio between the ray path x and the the averaged transit times (second term in equation 2). This term
corresponding transit time t. The result of this computation is (40 ms) is comparable with the sampling rate adopted to collect
shown in the following Eq. (1): times measurements (25 ms).
However, application of the standard error analysis sum-
 2  2
@v @v marised in Eq. (1) allows to take into account all systematic errors
s2v ¼ s2x þ s2t ; ð1Þ
@x @t made during the collection of the measurements, and by
considering the resulting standard deviation of the velocity,
and computing the partial derivatives we obtain: which is about 6% of the value of this observable, as an indicator
 2  x 2 of its precision, we can conclude that the resolution in the
1
s2v ¼ s2x þ 2 s2t : ð2Þ computation of the velocity field seems to be precise enough to
t t validate the applications it is used for in this work.
Eq. (2) represents the variance of the velocity field, where, in
order to evaluate st, we performed the weighted mean of the
transit times relating to the same ray paths with its standard 6. Discussion of the results
deviation (st), and as sx we used the mean value of 1 cm, which
represents the largest error in metric measurements. In this experimental study, the in-situ combined use of the low
In Eq (2), we neglected the correlation term because ray paths frequency ultrasonic and seismic methods and the integrated
x and transit times t are held by the equation: v=x/t. interpretation of their results have been effective in obtaining
In order to take into account the systematic errors that may quantitative information on the conservation state of the building
occur during collection of the transit time measurements, the materials of the ancient colonnade of the Church of San
standard deviation of the times in Eq. (2) was computed using a Bartolomeo in Central Sardinia from the knowledge of the elastic
weighted average technique, by taking as weights the squared status of the materials, while also checking the effectiveness of
inverse of the residuals between the transit time values and some restoration works.
their average. Owing to the nature of materials of the columns (a central
The procedure mentioned above was applied in following two body of squared pink trachite ashlar wrapped around by a
cases for each of the two columns analysed: concrete plaster bond covered with a cement lime mortar
brickwork), the ultrasonic tests were carried out by the surface
a) the direct ray paths from the transmitters to corresponding or indirect modality, i.e. placing transmitters and receivers on the
receivers, for example: T1–R1, if we follow the acquisition same side of the structure to check the conservation state of its
scheme of the longitudinal wave experiment shown in Fig. 5; superficial parts (brickwork and concrete).

Table 1
Estimation of the mean velocities error

2s (95%) Weighted mean times (ms) Distance (m) Mean velocities (m/s)

Undamaged column
(a) Direct paths 0.60+ /  0.04 1.70 + /  0.01 2816+ /  173
Percentage errors 6.7% 0.6% 6.1%
(b) Diagonal paths 0.64+ /  0.04 1.75 + /  0.01 2719+ /  159
Percentage errors 6.2% 0.6% 5.8%
Damaged column
(a) Direct paths 0.64+ /  0.04 1.70 + /  0.01 2679+ /  166
Percentage errors 6.2% 0.6% 6.2%
(b) Diagonal paths 0.69+ /  0.04 1.75 + /  0.01 2527+ /  160
Percentage errors 5.8% 0.6% 6.3%

Standard error analysis of the mean velocity field, for the undamaged column (first column on the right side of the plan) and the damaged column (first column on the left
side of the plan), (a) direct paths from transmitters to the corresponding receivers and (b) diagonal ray paths from transmitters to receivers set, 40 cm above and below the
receiver.
ARTICLE IN PRESS
68 S. Fais, G. Casula / NDT&E International 43 (2010) 62–69

The analysis of the results obtained with ultrasonic techniques propagation law to the expression of the mean velocity (Eq. (1)).
has generally shown that the propagation velocity values of The resulting RMS errors of the velocities relating to the altered
longitudinal pulses are representative of elastic conditions of the and unaltered parts of the analysed structure give very small
building materials that make up the shallow part of the columns, percentage errors of about 6%, which validates the measurements
and consequently, of their conservation state. To this purpose, it is themselves.
appropriate to point out that as deduced from tests carried out to
optimize the acquisition geometries of the ultrasonic data using
the step-by-step surface technique, the velocity values must be 7. Conclusions
interpreted as relative and not absolute. Nevertheless, though the
longitudinal velocity values measured by the ultrasonic technique The accuracy of the reconstruction of the longitudinal seismic
cannot be considered in the absolute sense, the velocity variations velocity distribution, anomaly localisation and resolution in the
detected in some sectors of the column faces have all been tomographic process can be improved by applying the proposed
accounted for by variations in the elastic conditions of the new experimental procedure based on the calculation of the
superficial materials (concrete and brickwork) and thus in their cross-correlation function (CCF).
mechanical behaviour. From an engineering standpoint, weak The interpretation of the cross-correlograms relating to several
areas, fissures, cracks and alterations in building materials are ancient columns of the investigated historical Church of San
potential structural defects. They often have low seismic velocity Bartolomeo and deduced from the previously described proces-
compared to intact materials [15,17]. sing sequence has been an objective constraint to precisely define
The ultrasonic longitudinal velocity map (Fig. 4) ranges the location and the geometry of anomaly zones of the transit
between 1500 m/s, which characterizes the sectors with shallow times of the longitudinal wave rays. Thanks to this knowledge,
altered zones, and the fastest velocity, 2100 m/s in correspon- together with the range velocity values deduced from different in
dence to a horizontal iron reinforcement within the concrete. In situ measurements (ultrasonic and SLWTT tests), it was possible to
the study context, since we were also investigating inside the implement and constrain the algebraic reconstruction technique
columns of around 1 m of diameter, thanks to the application of (ART) by introducing prior knowledge of the investigated
tomography with the seismic rather than the ultrasonic method medium.
by direct modality (transmitter and receiver on the opposite face From the analysis of the results of the SLWTT it was possible to
of the investigated object) better responses were obtained from locate damaged or altered zones inside the columns. These results
the deep materials. In fact the two methods mainly differ in the were confirmed at a reliability level greater than 90% by a few
use of longitudinal elastic waves of different frequencies and microcores on the most representative situations detected by
intensities. In this case, the seismic method operating at lower SLWTT. Moreover, we applied a standard error analysis to evaluate
frequency than the ultrasonic proved more effective. accuracy in the computation of the velocity field. The computed
The seismic longitudinal wave transmission tomography standard deviation of mean velocities gives a relative error of 6%
(SLWTT) allowed a detailed reconstruction of the velocity field of the computed mean velocity (about 2700 m/s). The RMS errors
through longitudinal sections inside the investigated columns. obtained by the previously described computation for the
The data processing based on the computation of the cross- velocities relating to the altered and unaltered parts of the
correlation function (CCF) improved the input velocity model for investigated structure give very small percentage errors (6%),
the tomographic reconstruction. which validated the results of the SLWTT measurements.
The data were divided into data sets and the CCF was
computed as previously described to be used as constraint to
Acknowledgements
the ART technique tomography [12,13] to insert prior knowledge
of the investigated medium. This subject was also discussed by
We sincerely thank Dr. Eng. R. Tocco for the helpful discussion
Roberts et al. (1989) [18]. From our experiment it can be seen that
on the application of the Seismic Longitudinal Wave Transmission
the CCF was characterised by low values in the undamaged
Tomography procedure.
material columns. Conversely, when a damaged material column
is investigated, the CCF values are close to the unit and can be
correlated with slow velocity parts of the building materials of the References
investigated structure (see Fig. 7 left). In particular, Fig. 7 (right)
shows the velocity map relating to damaged building materials of [1] Abbaneo S, Berra M, Binda L, Fatticcioni A. Non destructive evaluation of
bricks-masonry structures: calibration of sonic wave propagation procedures.
the column. Here the velocity ranges between 1800 m/s, char- In: Atti del international symposium non destructive testing in civil
acteristic of the most altered parts of the structure, and 4200 m/s engineering (NDT-CE), vol. 1, Berlin, Germany, 1995. p. 253–60.
relating to unaltered materials unambiguously revealing the [2] Binda L. Role of a Rilem committee: calibration of proposed test methods,
evaluation and strengthening of existing masonry structures. In: Atti del joint
shape and position of the anomalies. international workshop, Padova, 1997. p. 1–24.
The experimental procedure, which was applied to all the [3] Casula G, Fais S, Ligas P. Experimental application of 3-D terrestrial laser
columns of the Church of San Bartolomeo, affords a detailed scanning and acoustic tecniques in assessing the quality of stones used in
monumental structures. Int J Microstruct Mater Prop 2009;4(1):45–56.
reconstruction of the longitudinal velocity distribution in the [4] Concu G, Fais S. In the analysis of a viscous coupling agent effect in ultrasonic
investigated sections of the columns. measurements. In: Proceedings of the third international conference on NDT:
The application of this methodology allows us to distinguish organized by Hellenic Society of NDT (HSNT). NDT in antiquity and nowadays
skill application innovation, 15–18 October, Chania, Crete, Greece 2003. p.
between altered and unaltered materials inside a structure and 104–8.
their mechanical characteristics, and also to identify the most [5] Christaras B, Estimation of damage at the surface of stones using non-
significant damages in correspondence to some zones with destructive techniques. In: STREMAH 5 V. Fifth international congress of
structural studies, repairs and maintenance of historical buildings, San
anomalous elastic behaviour (Fig. 7 right). The position and
Sebastian. Southampton: Advances in Architectural Series of Computional
dimension of the damaged zones inside the column were verified Mechanics Publications; 1997. p. 121–8.
by direct investigation (microcores), whose strategic location was [6] Fais S, Ligas P, Palomba M, Tocco R. Evaluation of preservation state of
chosen based on the SLWTT results. monumental buildings by ND acoustic techniques and mineralogical studies.
In: Galan E, Zezza F, editors. Proceedings of the fifth international symposium
A test of significance analysis of the measurements of on protection and conservation of the cultural heritage of the Mediterranean
longitudinal velocity was implemented by applying the variance cities. Lisse: Swets & Zeitlinger; 2002. p. 307–14.
ARTICLE IN PRESS
S. Fais, G. Casula / NDT&E International 43 (2010) 62–69 69

[7] NORMAL—22/86. Misura di velocita di propagazione del suono. [12] Gordon R. A tutorial on ART (Algebraic Reconstruction Technique). IEEE Trans
In: Raccomandazioni NORMAL, CNR-ICR 1986. p. 1–7. Sci 1974;21:78–93.
[8] Fais S, Ligas P, Tocco R. Acoustic techniques for the evaluation of the building [13] Mazur EJ, Gordon R. Interpolative algebraic reconstruction techniques
materials of monumental structures. Geophysical Research Abstracts, Eur- without beam partitioning for computed tomography. Med Biol Eng Comput
opean Geosciences Union General Assembly Vienna, Austria, 24–29 April 1995;33:82–6.
2005. p. 1–3. [14] Zezza F. Evaluation criteria of the effectiveness of treatments by non
[9] Fais S, Casula G, Cuozzo G, Ligas P, Tocco R. Applicazione di metodologie destructive analysis. In: Proceedings of the second course of CUN University
geofisiche per la caratterizzazione delle rocce ornamentali. In: 281 Carrara School of Monument Conservation 1993. p. 198–207.
Marmotec, Carrara 30 Maggio–2 Giugno 2007. [15] Christaras B. Estimation of damage at the surface of stones using non-
[10] Galan E, Guerrero MA, Vasquez MA, Maert F. & Zezza F. Marble weathering: destructive techniques. In: STREMAH 5 V, fifth international congress of
relation between ultrasonic data and macroscopic observation. The case of the structural studies, repairs and maintenance of historical buildings, San
columns of the court of the lions at the Alhambra in Granada, Spain. Sebastian. Southampton: Advances in Architectural Series of Computational
In: Proceedings of the second international symposium, The conservation of Mechanics Publications; 1997 pp. 121–8.
monuments in the Mediterranean Basin, Geneve 1991. p. 193–207. [16] Taylor JR. An introduction to error analysis. The study of uncertainties
[11] Moropoulou A, Bakolas A, Aggelakopoulou E, Pineli T, Prassianakis I. in physical measurements, vol. 2. University Science Books Publisher; 1995
Extimation of elastic constants of stones, used in historic monuments, using p. 1–327.
ultrasonic technique and correlation to their microstructure parameters. [17] McKenzie CK, Stacey GP, Gladwin MT. Ultrasonic characteristics of a rock
In: Proceeding of third international conference on NDT: organized by mass. Int J Rock Mech Min Sci Geomech Abstr 1982;19:25–30.
Hellenic Society of NDT (HSNT). NDT in antiquity and nowadays skill [18] Roberts JDM, Belchamber RM, Lilley T, Betteridge D, Bishop I, Styles P. An
application innovation, Chania, Crete, Greece, Prassianakis, 15–18 October evaluation of computerized tomography for near-surface geophysical ex-
2003. p. 242–5. ploration. Comput Geosci 1989;5:727–37.