当前位置:首页 >> 机械/仪表 >>

Application of infrared thermography and geophysical methods for defect detection in architectural s


Engineering Failure Analysis 12 (2005) 875–892 www.elsevier.com/locate/engfailanal

Application of infrared thermography and geophysical methods for defect detection in architectura

l structures
Carosena Meola a,*, Rosa Di Maio b, Nicola Roberti c, Giovanni Maria Carlomagno a
Dipartimento di Energetica, Termouidodinamica Applicata e Condizionamenti Ambientali, ` Universita di Napoli Federico II, P.le Tecchio 80, 80125 Napoli, Italy b ` Dipartimento di Scienze Fisiche, Universita di Napoli Federico II, Complesso Universitario di Monte SantAngelo Via Cintia ed. G, 80126 Napoli, Italy ` Dipartimento di Scienze della Terra, Universita di Napoli Federico II, Largo S. Marcellino 10, 80138 Napoli, Italy Received 12 July 2004; accepted 27 December 2004 Available online 31 March 2005
a

c

Abstract The scope of the present study was a multi-methodological approach to non-destructive evaluation of architectural structures. Three dierent techniques such as infrared thermography, ultrasonics and electric-type geophysical methods were analysed to acquire information for a synergic use of the dierent methods, which may be useful for the estimation of the buildings degradation sources. The investigation was carried out in laboratory by considering specimens, which were made of a plaster layer over a support of marble, brick, or tu to simulate masonry structures. Air bubbles were intentionally created inside specimens to simulate detachments. Examples of in situ applications are also reported. In particular, infrared thermography was used for the inspection of the status of the tiles covering the walls of a building. 2005 Elsevier Ltd. All rights reserved.
Keywords: Architecture; Non-destructive evaluation; Infrared thermography; Ultrasonics; Micro-geophysics

1. Introduction Safety is becoming ever more a dominant concept of modern life involving at rst the building where human beings live and/or work. However, a safe home, or a safe oce, would essentially mean quality,
*

Corresponding author. Tel./fax: +39 081 7683389. E-mail address: carmeola@unina.it (C. Meola).

1350-6307/$ - see front matter 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.engfailanal.2004.12.030

876

C. Meola et al. / Engineering Failure Analysis 12 (2005) 875–892

high-quality materials, construction standards, maintenance. Of course, fabrication is a critical point since construction standards alone do not assure quality and do not prevent from deterioration of materials. The construction of a building should be accompanied by non-destructive evaluation of at least vital parts to ascertain for example the correct placement of grouted cells in concrete masonry unit (CMU) walls, or good adhesion of linings, or good performance of roofs and/or casings insulation, and so forth. In addition, periodic inspection should be made to evaluate existing conditions and discover building deciencies at an incipient stage to plan restoration and to prevent catastrophic failure. Often, we hear of unexpected building failure with a great loss of life and after it is discovered that the cause was water inltration. So that, it is important to use eective non-destructive techniques capable of giving valuable information about the main causes of deterioration of materials and structures used in the architectural eld. Generally, the degradation of architectural structures include: variations in concrete compaction and voiding, spalling or micro-cracking in masonry, reinforcement deterioration, mat size, concrete cover, cladding attachments, metal clamps and ties, wall ties, moisture content, leaks in roong systems, water inltrations. The aim of the present study was a multi-disciplinary and multi-methodological approach to nondestructive evaluation of architectural structures. Dierent techniques such as infrared thermography, ultrasonics and electric-type micro-geophysical methods were analysed. Experimental tests were carried out in laboratory by considering materials, which are used in the architectural eld. In particular, infrared thermography was also used for inspection of the status of tiles covering the walls of a building.

2. Description of specimens Three masonry specimens of dimensions 900 900 mm2 were fabricated. They included a support of marble, brick, or tu covered with plaster; defects, such as cork diskettes and air-lled plastic bags, were positioned between the support and the plaster to simulate detachments of the plaster from the support. The cork diskettes and the plastic bags were manufactured of three dierent diameters, d = 40, 60 and 100 mm; the thickness was approximately 1 mm for cork diskettes and 2 mm for plastic bags. In particular, the cork diskettes were waterproofed with specic paint to prevent them soaking water from the grout. A sketch of the marble support with defects glued over it is shown in Fig. 1; cork defects are indicated with a ‘‘C’’ while plastic defects are indicated with a ‘‘P’’. In the following, defects are shortly indicated with a ‘‘P’’, or a ‘‘C’’, followed by the diameter value (i.e., P100 for a plastic bag of diameter 100 mm, C60 for a cork diskette of diameter 60 mm and so forth). The plaster thickness, sp, was varied from 10 mm up to 55 mm. More specically, each specimen was divided into two parts: side A and side B. For side A, sp = 10 mm over both marble and tu and 25 mm over brick; for side B instead sp = 20 mm over both marble and tu and

Fig. 1. View of marble support and position of insets.

C. Meola et al. / Engineering Failure Analysis 12 (2005) 875–892

877

55 mm over brick. It is worth noting that the plaster thickness was uniform only over each side of the marble support specimen, while, due to fabrication diculties, a variation of about 30% can be observed over the other two supports.

3. Infrared thermography Infrared thermography (IRT) is a remote temperature mapping system which may be successfully exploited in many industrial and/or research elds, amongst others: meteorology, environment, medicine, architecture, engineering where the temperature (or surface temperature) represents a key parameter. Obviously, each eld presents specic characteristics and requirements which entail specic choice of infrared system, test procedure and data analysis. 3.1. Theoretical background IRT transforms the thermal energy, radiated from objects in the infrared band of the electromagnetic spectrum, into a visible image; each energy level is represented by a colour, or grey level. It basically includes a camera, equipped with a series of changeable lenses, and a computer. The core of the camera is the infrared detector, which absorbs the IR energy emitted by the object (whose surface temperature is to be measured) and converts it into electrical voltage or current. Any object emits energy proportional to its surface temperature. However, the energy really detected (by the infrared detector) depends on the emissivity of the surface under measurement and on the environment. In fact, a fraction may be either absorbed by the atmosphere between the object and the camera, either added as reected by the surface from the surroundings. To take into account these factors, calibration of the system by simulating real operating conditions has to be performed. The calibration function Ee eB=T A ; C 1

which recalls Plancks law relates the real amount of detected energy E to the emissivity e and to the surface temperature T of the object through the calibration constants A, B and C which take into account spurious quantities of energy from, or to, the environment. The old systems used a single detector, two dimensionality being achieved by rotating mirrors, or oscillating refractive elements (such as prisms) which scanned the eld of view (FOV) in both vertical and horizontal directions. The new generation FPA systems include a matrix of detectors to resolve the FOV. Measurements are generally performed in two dierent windows of the electromagnetic spectrum: shortwave (SW) (3–6 lm) and long-wave (LW) (8–12 lm). The origin of infrared thermography comes back to the early 1800s when William Herschel discovered thermal radiation, the invisible light later called infrared, but only in the mid-sixties IRT became a technique of temperature mapping. During the past several years such technique has evolved into a powerful investigative tool of non-destructive control. In fact, it is capable of providing useful information for the characterisation of materials and structures in a wide range of applications [1–13]. Basically, two thermographic techniques can be used for non-destructive evaluation: pulse thermography (PT) and lock-in thermography (LT). There are other techniques, which are variations of PT and LT since they include a dierent heating method, or a dierent processing algorithm. The pulse thermography is generally performed by stimulating the object by a heating pulse and monitoring its surface temperature evolution during the transient heating, or cooling, phase. Analysis with PT can be performed in two dierent modes: transmission and reection. In the transmission mode, the infrared camera views the rear face, i.e., opposite to the heating/cooling source. However, in real applications the

878

C. Meola et al. / Engineering Failure Analysis 12 (2005) 875–892

opposite side is often not accessible and so the reection mode must be applied with both infrared camera and stimulating source on the same side. The thermal energy propagates under the surface by conduction while the infrared camera monitors surface temperature variations. Obviously, the temperature distribution is uniform in case of uniform heating and homogeneous material, the presence of a defect at a certain layer interferes with the propagation of the thermal energy and causes a localised temperature dierence. Generally, a deep defect becomes visible later than a subsurface one and a large defect produces a larger temperature dierence than a smaller one. The evolution of the phenomenon can be observed by acquiring sequences of images, which, by means of postprocessing procedures, give information about size, depth and thermal resistance of defects [1,2,4]. Results may be aected by non-uniform heating and emissivity. The lock-in technique uses thermal waves [1–4,10,12,13]. Basically, the thermographic AGEMA Thermovision 900 Lock-In system (herein employed) is coherently coupled to a thermal wave source which is operated in such a way that a sinusoidal temperature modulation results; the modulation is generated by a non-linear electrical signal produced by the Lock-in Module which has a waveform table for this purpose. The heat source has to be calibrated (for each frequency) to ensure that the temperature waveform is really sinusoidal. The periodical transfer of heat at the surface (depth z = 0) of a homogeneous semi-innite material results in a (time dependent) thermal wave, which in one dimension, is given by z z T z; t T o exp exp i xt Az exp ixt /z; 2 l l where A(z) is the thermal amplitude, /(z) is the phase shift of the thermal wave travelling inside the material, l is the thermal diusion length which is calculated from the thermal diusivity a and the wave frequency f = x/2p r a : 3 l pf In the lock-in analysis the system collects a series of images and compares their temperatures computing amplitude and phase angle of the sinusoidal wave pattern at each point and so the resulting image may be an amplitude image or a phase image. The depth range, for the amplitude image, is given by l while the maximum depth, p, which can be inspected, for the phase image corresponds to 1.8 l. The LT technique, being insensitive to non-uniform heating and local emissivity variation, is particularly attractive for the investigation of mosaics, frescoes and paintings. A third technique is the pulse phase thermography (PPT) [7]; the surface is heated in pulse mode as in PT and results presented in terms of phase (amplitude) images as in LT. A comparison between PT, LT and PPT, for evaluation of frescoes, was made by Carlomagno and Meola [10]. 3.2. Experimental tests and data analysis Tests were performed in laboratory and in situ with both LT and PT methods and by using three thermographic systems: Agema 880LW, Agema 900LW coupled with the lock-in option and the SC3000 FPA. 3.2.1. Laboratory tests By employing the LT technique, it was observed that the plastic bag in the zone A4 (Fig. 1) of the specimen with marble support had been badly located; thus, it was removed with a portion of plaster and replaced again and covered with grout. Phase images of such zone before (f = 0.015 Hz) and after (f = 0.0037 Hz) restoration are shown in Fig. 2. The darker colour in Fig. 2(a) refers to the defect top and indicates an

C. Meola et al. / Engineering Failure Analysis 12 (2005) 875–892

879

Fig. 2. Phase images of the zone A4 for the specimen with marble support; (a) bad defect position f = 0.015 Hz; (b) restored f = 0.0037 Hz.

oblique position of the plastic bag. In Fig. 2(b), it is possible to see a central disk surrounded by an alone, which clearly shows the contour of the restored plaster. A probable explanation is that: the new added grout did not lled well the hollow caused by the plaster removal and so, during the grout solidication, a resizing occurred with formation of a fracture between the older and the new plaster; the new plaster did not have the same consistency of the previous one. However, what it is important for the purposes of the present paper is the capability of LT to detect the plastic bag underneath plaster and to discriminate between plaster layers manufactured in dierent times. Unfortunately, the heating frequency for the Agema 900LW Thermovision lock-in option can be varied in the range from 3.75 Hz (267 ms) down to 0.0037 Hz (273 s) in 15 intervals. The maximum depth for plaster (made of cement, sand and water) which can be inspected at the minimum selectable frequency is about 15 mm and so defects under a thicker plaster layer cannot be detected. A complete non-destructive evaluation of the three specimens was performed with pulse thermography. Sequences of thermal images were acquired with the FPA ThermaCam SC3000, which is able to detect very small temperature dierences (0.02 °C at 30 °C). Tests were carried out with the infrared camera positioned either far away enough to enclose in the eld of view all the three specimens, or close to the surface for detailed information over a small area. The specimens were vertically positioned with the camera viewing the plaster side. Thermal stimulation was performed with two lamps (1 kW each) when viewing the three specimens simultaneously and with one lamp when zooming at a smaller area; both lamps and camera were positioned on the same side (reection method). Sequences of thermal images were acquired during both transient heating and subsequent cooling at room temperature. From analysis of thermal images the following considerations can be made: defects visibility improved during cooling; defects under thick plaster became visible later (with time delay) with respect to those under thin plaster; for each specimen defects show better contrast on side A with respect to side B; the best defects visibility was displayed by the marble specimen and the worst by the tu specimen; plastic defects were better distinguishable than cork defects.

However, several factors inuence defects visibility: defects size and nature, depth, plaster consistency, type of support. In general, a large defect is better distinguishable than a smaller one. However, as it is evident from the 3D temperature map shown in Fig. 3, a small plastic bag is better visible than a large cork diskette. In fact, a P defect is characterised by a lighter plateau, while a C defect of the same diameter displays a more spiky appearance with a darker colour (lower temperature). Such behaviour can be explained in consideration of the smaller air volume really entrapped by cork, which certainly, notwithstanding the waterproong treatment, soaked a certain quantity of water from the grout during fabrication. A deep defect appears defocused because the thermal dispersion inside plaster becomes more important as the plaster

880

C. Meola et al. / Engineering Failure Analysis 12 (2005) 875–892

Fig. 3. Thermal image of the three specimens.

thickness increases. In addition, a compact plaster favours defect visibility while a porous plaster hinders defect visibility. In fact, the thermal conductivity of compact plaster is higher than that of air while the presence of lots of pores lled with air (porous plaster) reduces its thermal conductivity towards that of air; this eect contrasts the visibility of the defects which are just made of air. At last, also the thermal properties of the support could aect defects visibility. In fact, during cooling heat escapes also from the specimen backside (support) and higher the thermal conductivity of the support with respect to air (defects) better is the defects visibility and this justies the dierent behaviour of the two specimens with marble and tu supports. To show the capability of infrared thermography to detect the presence of humidity inside walls, water was let to ow between the thicker plaster (side B) and the tu support from the upside (specimen vertically positioned). Tests were carried out soon and after two days from water inltration. It was visualised a humidity stain in the upper left side; such stain was still visible after two days indicating a local water stagnation zone. All the aspects above described are summarised in the thermal image shown in Fig. 4. Such image was taken, by viewing simultaneously the three specimens, at the instant, during cooling, when almost all the detectable defects were visible. In particular, it is also visible (encircled in Fig. 4) the humidity stain after three days from water inltration. 3.2.2. Application in situ Infrared thermography was used to control the condition of the mosaics covering some external walls of the building shown in Fig. 5(a). The Agema 880LW system was used; the camera was rst positioned on the opposite foot-path far-away to view a large mosaic portion; after the camera-object distance was reduced to zoom into selected zones for details. Tests were performed in a sunny day. More specically, thermal images were acquired in the afternoon when the facade was in the shade, and so during transient cooling, after several hours to sun exposure. Results proved that some parts of the mosaics were in bad conditions

C. Meola et al. / Engineering Failure Analysis 12 (2005) 875–892

881

Fig. 4. 3D temperature distribution over side A for the specimen with marble support.

Fig. 5. Inspection of mosaics; (a) picture of the building facade; (b) thermal image of the zone inside the rectangle (a); (c) close-up view of the zone inside the rectangle of (b).

and were in want of restoration. A thermal image of the area inside the rectangle is shown in Fig. 5(b). High-temperature values in the zone inside the rectangle (Fig. 5(b)) indicate detachment of tiles; the air bubble between tiles and plaster acted as a heat barrier entailing a delay in cooling. A close-up view

882

C. Meola et al. / Engineering Failure Analysis 12 (2005) 875–892

(reduced camera–mosaic distance) of such zone is shown in Fig. 5(c). Restoration was promptly done to avoid accidents to the numerous persons who came to the building every day; the restorator conrmed the presence of large areas of detached tiles as indicated by the thermographic survey.

4. Ultrasonics The term ultrasonics is generally used to indicate the study and application of high-frequency sound waves (ultrasound), which are above the human hearing range (20 kHz). The history of ultrasonic technology came back to the discovery of the piezoelectric eect by Pierre Curie in 1880. One of the rst applications was sonar widely employed during World War II to detect enemy submarines. Recently, ultrasonic waves were exploited for cutting, cleaning, ow metering, non-destructive evaluation, medical imaging and sterilisation of surgical instruments, and so forth. Indeed, ultrasonic is the most popular non-destructive technique [14–17] since it is routinely used for control of industrial production in the aerospace eld where safe life is a dominated concept. 4.1. Basic principles The ultrasonic method is based on the property of sound waves to travel inside solid materials. A beam of ultrasonic energy is launched inside the material by exciting, with a high-voltage pulse, a piezo-electric crystal contained in a transducer in contact with the material; such transducer is called transmitter probe. Three basic types of stress waves are created: longitudinal, shear and surface waves. Generally, the propagation is the fastest for longitudinal waves and the slowest for surface waves. A local variation of material characteristics (acoustical impedance) aects the energy transmission; thus, the amount of energy that arrives at the receiver probe gives information about the material, or structure, characteristics (density, stiness, porosity). Measurements can be performed in three modes: direct, semi-direct, and indirect transmission. In the direct method, transmitter and receiver are placed on the opposite faces of the tested element. In the semi-direct method, the two transducers are placed at a 90° angle, whereas in the indirect method both transducers are placed on one face of the tested element. In the last mode a single probe can be used for transmitting and receiving. Note that a coupling liquid, or gel, must be used to assure good contact between probe and structure surface. The velocity of ultrasound waves depends on the elastic properties of the test material. In general, at a given pressure and temperature, the velocity, V, is constant for any wavelength, k, and any frequency, f V kf : 4 The frequency generally employed for non-destructive evaluation ranges from 100 kHz up to 50 MHz. Ultrasound is one of the most widely used techniques for non-destructive testing of aerospace structures [14–18]. The mostly used method is the pulse-echo performed with a single transmitter/receiver transducer of xed frequency in the range 1–10 MHz. This method measures the energy reected from internal defects; data are generally presented as C-scan images in the time domain. Image resolution is enhanced by using small, narrow-beam, high-frequency transducers. These observations generally apply for homogeneous materials. Unfortunately, concrete, being a heterogeneous mixture (cement, sand, aggregates), is not suited as conductor of stress waves and the interpretation of ultrasonic data is somewhat dicult because of the very low signal to noise ratio [19–22]. The main limiting factors in ultrasonic applications are the attenuative microstructure of concrete and the frequency-dependent attenuation characteristics. Attenuation is thought [21] to include three main contributions as: beam dispersion, absorption and diusion due to the presence of

C. Meola et al. / Engineering Failure Analysis 12 (2005) 875–892

883

grains, pores and cracks. Some researchers [19,22] analysed the attenuation characteristics of dierent types of cement-based materials and classied sand and mortar as the most attenuating elements. In particular, Landis and Shah [19] performed an experimental investigation with the point source/point receiver (PS/PR) ultrasonic technique [23] and found the attenuation to be a function of the degree of heterogeneity. Landis and Shah [19] observed, for each kind of material, an attenuation jump at a characteristic frequency, ranging from 900 kHz for pure cement down to 125 kHz for concrete. It has to be noted that the term concrete alone does not identify a specic material since several dierent kinds of concrete may be obtained by varying the type, amount and distribution of aggregate inclusions. Thus, it seems that ultrasonic measurements can provide only a qualitative picture of the concrete properties variations. 4.2. Test procedure and data analysis In the present study, the ultrasonic technique is used for non-destructive evaluation of specimens which belong to the family of cement-based materials; more specically such specimens simulated masonry structures being made of a support covered by plaster (Fig. 1). The plaster was a mixture of cement, sand and water, and so it was a kind of concrete; therefore in this section sometimes the plaster layer will be called ‘‘concrete’’. Tests were performed several months after fabrication when the plaster was completely dry; the specic weight of such plaster at the test time was equal to 2.58 g/cm3. Owing to the specimens geometry, two ultrasonic methods were employed: direct and indirect transmission. More specically, the direct method, which is also known as the longitudinal pulse velocity method (ASTM Standard C 597), was used to inspect the specimen with marble support. This choice was made on account that such specimen had both surfaces (plaster and marble) smooth enough for probes positioning; in fact, the support was obtained with four contiguous marble plates. The pulse velocity method relies on the measurements of arrival times and path lengths through the specimen; such method is also shortly called P-wave. Measurements were carried out with probes at sampling steps of 50 mm along x and y directions (Fig. 1). A couple of 150 kHz transducers were used. The two probes were positioned with care to establish good contact with the plaster surface and to align the two probes along the longitudinal wave pathway. In each point the arrival time, t, was measured and the velocity Vp through such structure plaster + marble was calculated by the relationship s 5 Vp ; t where s is the overall thickness (plaster sp + support ss). A velocity map is shown in Fig. 6. The rst observation is that the thinner side A shows velocity values higher than side B. This observation complies with expectation because the pulse travels partly in plaster and partly in marble and, being the pulse velocity higher in marble than in concrete, the attenuation is proportional to the concrete thickness. Another observation is that both sides are characterised by an almost uniform background with local maxima and minima. Such maxima and minima should, respectively, indicate local stiness and voids, but, minima do not always coincide with the position of defects. In fact, in side B the two minima located at x = 55 cm, y = 45 cm and x = 65 cm, y = 23 cm indicate the presence of the P60 and P100 defects (Fig. 1); the other minimum located at x = 60 cm, y = 65 cm should indicate the C100 defect, but its position is not correctly identied. No signs of the C60, C40 and P40 defects can be recognised. In side A the minimum on the bottom could indicate the P100 defect but its position is not correctly identied. However, the chosen test structure is very complex for ultrasonic inspection. The main disadvantages are attributable to a poor contrast between the elastic characteristics of the defects and the concrete, and to the strong attenuation of the sound through the concrete. The C defects contain a very low volume of air

884

C. Meola et al. / Engineering Failure Analysis 12 (2005) 875–892

A 80
y[cm]

B
2800 2600 2400

60

2200 2000

40

1800 1600 1400

20

1200 1000

0 0

800

20

40

60 x[cm] 80

V p [m/s]

Fig. 6. Velocity map of the specimen with marble support.

because, notwithstanding the waterproong treatment, the cork could have soaked water from the grout during fabrication reducing the air volume initially present in it. Thus, such defects may be confused with the numerous diusers (grains and pores) present in the concrete. In the case of plastic bags, a velocity rise should be observed in the passage from plaster to plastic but such rise could be compensated by a decrease in the air layer. Again, the velocity dierences could be not signicant for clear defects identication. Another adverse eect may be the further wave attenuation in the ssures, i.e., voids, inevitably present between two contiguous marble plates. The indirect method was used for the analysis of the specimen with brick support. Indeed, the rear surface of such specimen was not suitable for probes positioning because of three main problems: the presence of a bricks protective net, a surface not regularly smoothed and the large number of brick blocks that caused many ssures between contiguous blocks. With the indirect method the pulse velocity can be calculated by referring to the surface wave. It is the least sensitive arrangement for velocity calculation, but the more appropriate for the quality analysis of shallower layers. Therefore, the ultrasonic measurements performed on the brick specimen have been devoted only to the analysis of the plaster layer. Measurements were carried out on the thicker side B with the tandem technique by moving both transmitter and receiver; a frequency of 220 kHz was used. We remember that the brick specimen had the thickest plaster layer: 2.5 cm on side A and 5.5 cm on side B. The main results coming from this survey can be synthesised as follows: a constant wave travelling time, ts, when both transducers were located on the sound material; a delay, or lack of signal, when, respectively, one transducer, or both, were positioned in correspondence of the plastic defects; no signicant inuence of the cork diskettes on the elastic waves pattern. Such observations were summarised in the three plots of Fig. 7. To facilitate data interpretation, the t measured values on the whole panel were normalised with respect to the characteristic time, ts, of the sound material. In particular, when both transducers were over the P100 defect (Fig. 7(a)) no signal was observed, probably for the wave high attenuation induced by the air inside the plastic bag. In fact, in correspondence of the smaller defects (i.e., P60) the attenuation eect is lower and the measurement became reliable (Fig. 7(b)).

C. Meola et al. / Engineering Failure Analysis 12 (2005) 875–892

885

Fig. 7. Normalised wave travelling time distribution over side B of the brick specimen; (a) across P100; (b) across P60 and C60; (c) across C100.

886

C. Meola et al. / Engineering Failure Analysis 12 (2005) 875–892

5. Electric-type geophysical methods The non-destructive geophysical prospecting methods are able to reconstruct detailed physical and geometrical images of hidden volumes from surface measurements. In the last 20 years, they were successfully applied in the cultural heritage (CH) eld, implying the reconstruction of buried archaeological remnants, and recently more specically employed in the inspection of architectural structures with the aim to dene the endogenetic degradation sources of buildings. In particular, the geophysical study we present basically consists of application and integration of electric-type geophysical methods: self-potential and geoelectrics. For both methodologies the physical parameter we measure is the electric potential drop whose source is, respectively, any buried stationary electric charge conguration inside the investigated volume, and a continuous current injected into the body by an electrodes couple put on its surface. 5.1. Basic principles 5.1.1. The self-potential method The self-potential (SP) method is used to reveal natural potential drops existing in any material, very likely caused by electrokinetic eects related to in-body uids circulation. These potentials are observed at the ground surface at the end of a passive line in which impolarizable electrodes are grounded. The aim is to obtain a SP contour map in order to evidence anomalous zones likely connected with underground anomalous electric charge concentrations, sustained by polarisation mechanisms due to the electrolytic solutions movement. This means that the application of the SP method at the micro-scale may help to identify anomalies related to water inltration through stones and similar architectural block devices, in the study of the internal causes of buildings deterioration. This approach would have perhaps never been utilised in the past, essentially because of the lack of a reference phenomenological model and of a suitable 3D tomographic imaging of the charge polarisation centres. Recently, by a conceptually simple adaptation of the water circulation/charge polarisation eld phenomenology [24] to the scale of stones, columns and walls, using the newly developed data acquisition techniques and data tomographic processing methods [25,26], have been put in evidence SP micro-anomalies, which have opened new vistas in the conservation state diagnosis of any architectural structure [27]. In particular, the 3D tomography method is based on the concept of charge occurrence probability (COP). Briey, the method founds on the decomposition of a synthetic SP surface eld into a sum of elementary contributions due to a discretised distribution of charge accumulation centres. The inversion problem consists in recovering the most probable discretised charge distribution underground, responsible of the measured SP eld, through a normalised cross-correlation procedure between the observed natural electric eld and a theoretical scanner function. This procedure properly denes the COP function. Using this physical scheme, the tomographic process consists of scanning the whole lower half-space (the tomospace) along a sequence of horizontal planes parallel to the eld plane, for any xed depth. For each depth, the complete grid of regularly distributed COP values in the horizontal plane can be utilised to draw contoured slices in order to single out the zones of highest probability of concentrations of polarised electric charges. 5.1.2. The geoelectrical method The direct current geoelectrical (DCG) method consists in the experimental determination of a set of apparent resistivity values, which, according to mathematically very complex composition rules, characterise any inhomogeneous arrangement of physically dierent materials in electrical contact. By appropriate interpretation procedures it is then possible to recover from the experimental data set the true resistivity and the geometry of each constituent material.

C. Meola et al. / Engineering Failure Analysis 12 (2005) 875–892

887

The standard eld procedure consists in the determination of an as dense as possible set of apparent resistivity values in correspondence of as many dierent positions of the electrode device, which is used to generate the input current eld and to detect the electrical response of the underground materials stimulated by the current. In particular, for near surface prospecting, due to the generally limited dimensions of the targets, ultra-high resolution (u.h.r.) techniques are needed. To this purpose, the geoelectrical pseudosection technique along a selected prole can lead to an extremely detailed picture of the electric resistivity behaviour across the vertical plane through the prole. The true resistivity distribution is, then, derived as solution of the so-called inverse problem, which is in principle unique in absence of disturbances of any kind. The algorithm commonly used for the tomographic inversion of the resistivity pseudosections refers to a least-squares deconvolution method [28], which is based on a linearisation process of the apparent resistivity values as function of the involved real resistivity. Resistivity in materials depends on many factors; the most important is the presence of ionised water and/or metal-like mineral particles in the pore structure, which entail conductive internal paths, in contrast with resistive situations characterised by compact rock matrices and/or dry pores. In general, in archaeological research, the presence of a high geoelectrical anomaly is a good indicator of some resistive structures, such as cavities, stone walls or foundations, hosted within an otherwise less resistive material due to presence of circulating uids. In the case of buildings, the reverse situation becomes the interesting case: the capillarity ascent of humidity and sometimes the ingression of more or less aggressive waters may provoke internal alteration nuclei and become the hidden, potential endogenetic source of degradation or even disgregation of stone chunks and/or surface plasterings. 5.2. Experimental tests and results 5.2.1. SP measurements SP drops were measured between contiguous points 25 mm pitch along the y axis; y-proles were taken at steps of 44 mm along the x axis as shown in Fig. 8. Measurements were performed with electrodes which are commonly used in the medical eld and which allowed overcoming some of the problems (high-contact resistance, signal instability, etc.) of conventional electrodes (copper plates, needles, etc.). The surface of each specimen was wetted with about 2000 cm3 of tapped water per m2.

44
y [mm]

M 25 N

x [mm]
Fig. 8. Sketch of the micro-geophysical survey station sites.

888

C. Meola et al. / Engineering Failure Analysis 12 (2005) 875–892

Fig. 9 shows the SP maps (slides on the gure top) and the tomographic sequences for the three analysed specimens. As we can see, low SP values (of the order of few mV) generally characterise the hosting materials, while large SP variations are observed in correspondence of the defects location, very likely ascribable to polarisation charge concentrations originate by water accumulation over the defects surface. This correspondence is particularly evident for the specimen with marble support, whereas the more uniform SP values distribution characterising the specimen with brick support aects the defects visibility. Finally, as regards the specimen with tu support, relatively high SP anomalies are observed in side A while side B shows a quite uniform SP drops distribution; a possible explanation is the presence of pores in side B, responsible of a more homogeneous water disposal which aected the defects visibility. Such evidences seem all ascribable to the plaster and support porosity level. In particular, the presence in the plaster of pores and/or micro-cracks should have favoured water inltration, responsible of ionic charge concentrations at the boundaries of any material discontinuity. Moreover, a too-high porosity level should have entailed a too-much water absorption giving rise to a uniform charge distribution, which has masked the presence of the defects. Such eect strengthens if also the support is porous.

y[cm]

A

B

y[cm]

B

A

y[cm]

A

B

x[cm]
-70 -50 -30 -10 10 30 50 -16 -12 -8 -4 0 4

x[cm]
8 -50 -30 -10 10 30

x[cm]
50

SP drops [mV]
y[cm]
y[cm]

SP drops [mV]
y[cm]

SP drops [mV]

1_
y[cm]

1_
y[cm]

1_
y[cm]

1.5 _
y[cm]

2_
y[cm]

1.5 _
y[cm]

depth [cm]

2_

3_
y[cm]
y[cm]

2_

depth [cm]

depth [cm]
2.5 _ 3_ 3.5 _
x[cm]

y[cm]

4_
y[cm]
y[cm] y[cm]

2.5 _

3_
y[cm]

5_
y[cm]

y[cm]

3.5 _
x[cm]
-0.24 -0.16 -0.08 0.00 0.08 0.12 0.16 0.20

6_

x[cm]
0.35 -0.15 -0.10 -0. 05 0.00 0.05 0.10 0.14 0.1 6

-0.19 -0.12 -0.07 0.00

0.07

0.18

0.29

charge occurrence probability values

charge occurrence probability values

charge occurrence probability values

a) marble

b) brick

c) tuff

Fig. 9. SP drop maps (gure top) and tomographic images at several depths of charge occurrence probability values. (For interpretation of the references to colour in this gure legend, the reader is referred to the web version of this article.)

C. Meola et al. / Engineering Failure Analysis 12 (2005) 875–892

889

In the light of the above considerations, the better defects visibility of the specimen with the marble support is attributable to a poor porosity of the plaster layer (already visible on the surface) and to the compact structure of the support; instead, the weak and partial defects visibility, respectively, observed for brick and tu specimens, can be ascribed to the presence of a large number of interconnected pores and/or microcracks net in the plaster layer and to the porous nature of the supports. A validation of our hypotheses is provided by the results of the 3D tomographic inversion (Fig. 9) of the SP data observed for the three specimens. The tomographies refer to images between 10 and 35 mm below the surface level for the marble (a) and tu (c) specimens, and between 10 and 60 mm for the brick specimen (b). It is worth noting that positive values indicate positive charge accumulations, while negative values indicate negative charge concentrations. As can be seen, the positive (red in web version) and negative (blue in web version) nuclei are observed in both sides of the marble and brick specimens, while for the tu specimen only the side A shows the presence of polarisation centres, indeed the B sector exhibits a quite uniform and very low intensity COP values distribution (green area in web version). The charge accumulations for the marble and tu (side A) specimens attain their highest absolute COP values at the depth of 20 mm below surface level, for both positive and negative nuclei. The location in depth of these clusters coincides rather well with that of some defects (Fig. 1). A singular behaviour, instead, is shown by the tomographic sequence of the brick specimen. In fact, contrary to expectations, high COP values are observed on the thicker side (B), i.e., with lower water content, at the depth of the defects location, but not in perfect correspondence with them, as already outlined by the relative SP map. According to our previous hypothesis, the behaviour of the tu side B could be ascribed to both plaster and support high porosity that favoured an uniform water distribution in the whole system (plaster plus tu). Thus, such system acted as a homogeneous material precluding the defects visibility, also in consideration of their very low thickness ($1 mm). Moreover, a high porosity level and/or a large micro-fracture net of the plaster covering the brick side B could have favoured a major water inltration with respect to the side A causing charge accumulation processes. Also in this case, a not clear defects location seems imputable to the high porosity characterising the plaster and the support. 5.2.2. DCG measurements To verify the occurrence of a signicant uid circulation in the deep layers, as suggested by the results of the SP surveys, ultra-high resolution geoelectrical tomographies have been carried out for the three specimens along the same SP measurement proles (full lines in Fig. 8). The results of the resistivity pseudosections inversion, performed with the Loke and Barker algorithm [28], are shown in Fig. 10. In particular, such gure shows a 3D representation of the resistivity values obtained by interpolating the data extracted at three depths from the inverted resistivity pseudosections. Generally, the three images (Fig. 10) exhibit: Non-homogeneous distribution of the resistivity values inside plaster and support. This can be ascribed to a non-uniform water distribution induced by local porosity variations, as already outlined by the SP prospecting. Higher resistivity values on the side B respect to the side A. Such behaviour is due to the less water content in the thicker side (B). In fact, both sides have been wetted with the same water volume (1 l). A not good defects detection, above all for the smallest ones. This can be imputable to the used too-large space interval (25 mm), with respect to the defects size and/or depth. In this context, it is worth to point out that the main disadvantages of the used electrodes is linked to the dimension (about 25 mm of diameter) of the gel sponge around the metal electrode. In fact, while on the one hand such size ensures a good contact with the structure surface, on the other hand, it does not permit the use of a sampling step smaller than 25 mm. Consequently, it not possible to outline very small and/or low depth defects.

890
A

C. Meola et al. / Engineering Failure Analysis 12 (2005) 875–892
B
y[cm]

y[cm]

B

A

y[cm]

A

B

1-

x[cm]

1y[cm]

1x[cm]

x[cm]

depth [cm]

y[cm]

y[cm]

2y[cm]

x[cm]

4y[cm]

2x[cm]

x[cm]

y[cm]

3-

x[cm]

63.4 1.6 2.2 2.8 3.4 4.0 4.6

3x[cm]

x[cm]

1.0

1.4

1.8

2.1

2.5

2.8

3.1

log resistivity [m]

5.2

6.0

1.0

1.5

2.0

(a) marble

log resistivity [m]

log resistivity [m]

2.5

3.0

3.8

4.6

5.2

(b) brick

(c) tuff

Fig. 10. 3D tomographic representations of the electrical resistivity values.

Looking at each image specically, it can be observed: The presence of an area characterised by high-resistivity values on the bottom of the marble specimen (side A); such area corresponds to the restored zone in which the plaster being more compact absorbed less water. This result well agrees with the thermographic observation. Two low resistivity anomalies in the tu specimen (Fig. 10(c)). The rst is a strip in the side A at the depth of 2 cm below surface level, which starts approximately at y = 45 cm and goes transversally up to y = 55 cm. Its shape seems well dene a transversal fracture which entrapped enough water to evidence its presence. The second one is visible on the left top in the side B from 2 to 3 cm of depth. Its circular geometry let to suppose the presence of a cavity on the tu support that caused a local water stagnation so as evidenced by the thermographic investigation. At last, another observation coming from the DCG survey and that, in our opinion, is the most relevant nding, is the discrimination between the dierent supports and the support and the plaster. In fact, the resistivity ranges characterising the specimens at the depth of the supports are: 103–104 X m for the brick, 102–103 X m for the tu and of the order to 102 X m for the marble support. The observed resistivity intervals fully agree with those reported in the literature [29] for these materials in wet conditions. Moreover, for all the investigated specimens, the resistivity values characterising the plaster are always less than 102 X m, i.e., lower (at least of one order of magnitude) than the three support resistivity values. Thus, such technique is able to recognise the eventual contact between dierent materials.

6. Concluding remarks From a comparison of the results obtained with the dierent methods it can be inferred that: Infrared thermography is the easiest and quickest technique for detection of defects in cement-based materials. In particular, LT is capable of supplying detailed information about size, position and nature of defects and of discriminating between layered structures and local variations of concrete consistency. Unfortunately, LT is limited to thin concrete layers. The PT technique allows for a more in-depth

C. Meola et al. / Engineering Failure Analysis 12 (2005) 875–892

891

analysis since it was capable of detecting, with good contrast, defects which were very thin (1 mm), quite deep (55 mm) and of thermal characteristics close to the hosting material. The thermal images brought also evidence of the inuence of the type of support on the defects visibility. However, it is very dicult to identify with PT the interface between stratied materials. In fact, the same signal degradation may be induced by the support (specic thermal characteristics), or by the defect (thermal characteristics too close to the sound hosting material, or too small, or too deep). The ultrasonic technique, unlike infrared thermography, cannot supply in a fast way detailed information about size and location of small defects in thin layers of concrete. Instead it can provide, through the attenuation of the P-wave, information useful for the evaluation of structural disomogeneities also in presence of very thick materials. The electric-type geophysical methods seem not properly adequate for detection of small and near-surface defects. Conversely, through 3D representations of the natural electric charge accumulations (SP) and of the electrical resistivity (DCG), it can be established the porosity level and the distribution of voids and micro-cracks in the whole masonry structure (plaster + support). In particular, the highest porosity level was found on side B of the specimen with tu support; this nding justied the worst defect visibility observed over such side in the thermal image (Fig. 4(a)). From the resistivity values distribution inside the investigated volumes, it was possible to identify the three dierent kinds of support. In synthesis, we can conclude that a combined use of the proposed techniques may be helpful for nondestructive evaluation of architectural structures. Infrared thermography, as a remote imaging system, can be used for quick evaluation of large surfaces; what is more, being a non-contact technique, it can be used also for the study of precious artworks such as mosaics, frescoes, paintings, etc. The ultrasonic method can be used to analyse the in-depth conditions, such as the thickness of the walls and the characteristics of the support under the plaster. The electric-type geophysical methods are helpful for high-depth evaluation. In addition, such methods can be used, in conjunction with infrared thermography, for material characterisation at low-depth. Thus, within a combined use of the three dierent methodologies, it seems possible to outline and characterise the endogenetic degradation sources from micro- to macro-scales.

Acknowledgement The technical assistance of Mr. Giuseppe Sicardi (DETEC) is acknowledged.

References
[1] Annual SPIE thermosense conferences, Bellingham, WA; 1978–2002. [2] Biennial quantitative infrared thermography (QIRT) conference; 1992–2002. [3] Busse G, Wu D, Karpen W. Thermal wave imaging with phase sensitive modulated thermography. J Appl Phys 1992;71(8):3962–5. [4] Vavilov V. Thermal non-destructive testing: short history and state-of-art. In: Balageas D, Busse G, Carlomagno GM, editors. Proceedings of the Qirt92 conference, Chatenay-Malabry, France; 1992. p. 179–93. [5] Feit E. Infrared thermography saves shelburne middle school time and money. Mater Eval 1987;45:400–1. [6] Grinzato E, Bison PG, Marinetti S, Vavilov V. Non-destructive evaluation of delaminations in fresco plaster using transient infrared thermography. Res Nondestr Eval 1994;5:257–71. [7] Maldague X, Marinetti S. Pulse phase infrared thermography. J Appl Phys 1996;79:2694–8. [8] Grinzato E, Bison PG, Bressan C, Mazzoldi A. NDE of frescoes by infrared thermography and lateral heating. In: Balageas D, Busse G, Carlomagno GM, editors. Proceedings of Qirt98, Lodz, Poland; 1998. p. 64–70. [9] Carlomagno GM, Meola C. Infrared thermography in the restoration of cultural properties. In: Proceedings of SPIE thermosense XXIII conference Orlando, USA, vol. 4360. 2001. p. 203–16.

892

C. Meola et al. / Engineering Failure Analysis 12 (2005) 875–892

[10] Carlomagno GM, Meola C. Comparison between thermographic techniques for frescoes NDT. NDT&E Int 2002;35(8):559–65. [11] Grinzato E, Bison PG, Marinetti S. Monitoring of ancient buildings by the thermal method. J Cultural Heritage 2002;3:21–9. [12] Meola C, Carlomagno GM, Squillace A, Giorleo G. Non-destructive control of industrial materials by means of lock-in thermography. Meas Sci Technol 2002;13(10):1583–90. [13] Meola C, Carlomagno GM, Giorleo L. Non-destructive evaluation of bonded structures with lock-in thermography. J Adhes Sci Technol 2003;17(9):1207–22. [14] Blitz J, Simpson G. Ultrasonic methods of nondestructive testing. London: Chapman & Hall; 1996, ISBN 0 412 60470 1. [15] Chang Y-F, Chen S-C. Imaging hilbert-transformed ultrasonic data. Res Nond Eval 2001:97–103. [16] Krautkramer J, Krautkramer H. Ultrasonic testing of materilas. 4th ed.. Berlin: Springer; 1990, ISBN 0-387-51231-4. [17] Lew Yan Voon LFC, Bolland P, Laligant O, Gorria P, Gremillet B, Pillet L. Gradient-based hough transform for the detection and characterization of defects during non-destructive inspection. In: Proceedings of the SPIE IS&T/SPIE symposium on electronic imaging science and technology San Jose, California, USA, vol. 3029; 1997. p. 140–46. [18] Huang YD, Fryen L, Wevers M. Quality control and nondestructive tests in metal matrix composites. J Nondestruct Eval 2001;20(3):113–32. [19] Landis EN, Shah SP. Frequency-dependent stress wave attenuation in cement-based materials. J Eng Mech ASCE 1995:737–43. [20] Smolarkiewicz PP, Nogueira CL, Willam KJ. Ultrasonic evaluation of damage in heterogeneous concrete materials European congress on computational methods in applied science and engineering ECCOMAS 2000, Barcelona; September 2000. [21] Garnier V, Corneloup G, Toppani E, Leygonie M. Non-destructive evaluation of concrete damage by ultrasounds 15th world conference on non-destructive testing (WCNDT) Rome, Italy; 2000. [22] Gaydecki PA, Burdekin FM, John DG. Propagation and attenuation of medium-frequency ultrasonic waves in concrete. A signal analytical approach. Meas Sci Technol 1992;3(1):126–34. [23] Sachse W, Kim KY. Point source/point receiver materials testing. In: Thompson DO, Chimenti DE, editors. Proceedings of review of progress in quantitative nondestructive evaluation conference. New York: Plenum Press; 1986. p. 311–20. [24] Di Maio R, Patella D. Self-potential anomaly generation in volcanic areas. The Mt.Etna case-history. Acta Vulcanologica 1994;4:119–24. [25] Patella D. Introduction to ground surface self-potential tomography. Geophys Prospect 1997;45:653–81. [26] Patella D, Piro S, Versino L, Cammarano F, Di Maio R, Mauriello P. The IGAPS project: integrated geophysical acquisition and processing system for cultural heritage. In: Pasquinucci M, Trement F, editors. Non-destructive techniques applied to landscape archaeology. Barker G, Mattingly D, editors. The archaelogy of mediterranean landscapes series. Park End Place, Oxford: Oxbow Books; 2000. p. 96–103. [27] Carlomagno GM, Di Maio R, Meola C, Roberti N, Curciotti R. Metodologie geosiche e termograche per la salvaguardia di Beni Architettonici Paesaggio Urbano. Maggioli Ed 2003;2:32–5. [28] Loke MH, Barker RD. Least-squares deconvolution of apparent resistivity pseudosections. Geophysics 1995;60:1682–90. [29] Telford WM, Geldart LP, Sheri RE. Applied geophysics. Cambridge: Cambridge University Press; 1996. p. 755.


相关文章:
Photovoltaic systems thermography from the air using drones
Photovoltaic systems thermography from the air ...The device was to have a high infrared ...serious defects in photovoltaic systems from the ...
更多相关标签:
thermography | geophysical | geophysical research | ion geophysical | geophysical survey | detestation | www.nanjingdete.com | nanjingdete.com |