Quality Control

NDTCE’09, Non-Destructive Testing in Civil Engineering Nantes, France, June 30th – July 3rd, 2009 Calibration and on-site application of radar and sonic methods for quality control of reinforced masonry Diego AROSIO1, Francesca DA PORTO2, Flavio MOSELE2, Stefano MUNDA1, Luigi ZANZI1 1 2 Politecnico di Milano, Milan, Italy, luigi.zanzi@polimi.it University of Padova, Padova, Italy Abstract In the framework of the DISWall research project, funded by the European Commission, quality assessment procedures for modern reinforced masonry buildings, involving nondestructive test methods, were performed. Different testing techniques were applied on two reinforced masonry systems, based on concentrated vertical reinforcement and on the use of perforated clay units. This contribution focuses on results obtained by radar and sonic techniques, applied in laboratory on two masonry specimens which were built including known defects in the masonry. The tests were also carried out on-site, on real walls of a selected case study. Among the various techniques, dynamic tests were also performed. Their results are reported in Mosele, da Porto, Modena (2008), of the present conference. Résumé Dans le cadre du projet de recherche DISWall, financé par la Commission Européenne, ont été mises en œuvre des procédures de contrôle de la qualité de la construction moderne en maçonnerie armée, fondées sur l'usage des techniques non-destructives. Diverses techniques d’épreuve ont été appliquées sur deux systèmes différents en maçonnerie armée : avec renforcement vertical ciblé, et, basé sur l'usage de blocs de briques creuses. Sont présentés dans cette contribution, les résultats des méthodes radar et soniques appliquées dans le laboratoire sur deux spécimens de maçonnerie construits avec une série de défauts. Les mêmes méthodes ont été répétées in-situ, sur les panneaux de maçonnerie armée d’un bâtiment. Parmi les différentes techniques appliquées, des tests dynamiques ont également été effectués. Les résultats sont présentés dans Mosele, da Porto, Modena (2008), à cette conférence. Keywords High-frequency radar, sonic tests, signal processing, diagnosis of defects. 1 Introduction Reinforced load-bearing masonry walls can be very effective in improving the seismic resistance of buildings [1], providing at the same time satisfactory internal environment. Nevertheless, use of complex construction technologies, employment of low workmanship and improper construction practices can lead to defects that completely alter the behaviour of walls, and make useless the effects of reinforcement. Moreover, the lack of procedures for quality control makes it impossible to define and evaluate standard parameters against which measuring the quality of the final product, i.e., the masonry wall. To solve these problems, in the framework of the DISWall project, various innovative systems for reinforced masonry walls were developed and were subjected to extensive experimental and numerical studies. In particular, two systems based on the use of perforated clay units were studied in Italy. One consists of horizontally perforated clay units with recesses for the placement of the horizontal reinforcement, which is situated each other bed joint and can be made of either steel bars or NDTCE’09, Non-Destructive Testing in Civil Engineering Nantes, France, June 30th – July 3rd, 2009 prefabricated steel trusses. The vertical reinforcement is positioned into concentrated confining columns, made with vertically perforated clay units. The thickness of the units as well as of the walls is about 300 mm (Figure 1 left). The second building system entails the use of two vertically perforated clay units: ‘H’ shaped units alternate with ‘C’ shaped units, which can be put in place after the vertical reinforcement has been already placed (Figure 1 right). Several ND techniques were used to examine the abovementioned building structures. Among these, Ground Penetrating Radar (GPR) and sonic tests were performed on laboratory specimens made with the first construction system and on real walls built with the second construction system. These methods have been widely employed for assessing reinforced concrete structures such as civil infrastructures [2] and have been adapted and experimentally applied also on historic masonry buildings [3]. However, GPR and sonic methods have not been extensively applied, so far, to modern masonry buildings. The main aim of the testing campaign was that of evaluating the effectiveness of the proposed non-destructive techniques for localizing typical defects of reinforced masonry walls; calibration measurements on laboratory specimens, were followed by on-site applications to validate the results. Figure 1. 2 System with horizontally perforated units (left); with ‘H’ and ‘C’ units (right). Laboratory tests Two specimens 1,7 m high and 2,2 m large were erected according to the first construction system represented in Figure 1 (left). Both specimens were built with horizontal reinforcement placed each other bed joint, made either with two steel rebars or with semiprefabricated trusses. Each specimen also included vertical reinforcements placed in two confining columns. One of the specimens had properly done columns, with bars overlapping at the bottom of the specimen (A1-11 to 41 and A1-16 to 46, Figure 3 left), but a series of defects concentrated in the bed joints, such as the absence of mortar filling between the horizontal rebars (defect α, A1-11 to 16 and A1-71 to 76, Figure 3 left) and the absence of cover around the bars (defect β, A1-31 to 36 and A1-51 to 56, Figure 3 left). The other specimen was characterized by properly done bed joints (besides some bars having “partial cover”, defect γ, A1-31 to 36 and A1-51 to 56, Figure 3 right), but a series of defects concentrated in the vertical confining columns, e.g. complete lack of mortar filling and absence (A1-11 to 41 and A1-16 to 46, Figure 3 right) or presence (A1-41 to 71 and A1-46 to 76, Figure 3 right) of vertical rebars. Furthermore, some damage was caused to the upper course (the 7th) of units of both specimens, to check the effectiveness of NDTs in identifying damaged areas. More details on the laboratory specimens are available in Zanzi et al. [4]. 2.1 GPR application The GPR tests were performed with a very high frequency, dual polarized antenna (2GHz), on the specimens as they were built and also after them being plastered. Data were collected along horizontal and vertical profiles (with both 2D and 3D surveys) intercepting all NDTCE’09, Non-Destructive Testing in Civil Engineering Nantes, France, June 30th – July 3rd, 2009 the remarkable features and defects affecting vertical and horizontal reinforcements. Data quality was exceptionally good; high resolution allowed to detect singular diffractions from each hole of the clay units and powerful reflections from bed-joints with and without reinforcements could be observed easily. Vertical profiles were accurately examined and radar signatures were compared in order to assess different conditions of the bed-joints. Data processing revealed the possibility to differentiate the absence of defects, the lack of mortar covering the rebars, plus the absence of reinforcements. Partial mortar cover and lack of filling between steel bars could not be classified instead. Figure 2 left summarizes the results, illustrating representative signatures of the three situations that could be discriminated for the steel bar case. Similar conclusions were drawn for the steel truss case; plastered specimens gave comparable results. Consequently, a cross-correlation algorithm for an automatic analysis and classification of the bed-joints on the basis of the radar response was developed. The output of this automatic approach was very encouraging since a classification success rate very close to 100% was observed. An assessment analogous to the previous one was then carried out for the vertical reinforcements. It was concluded that the area of bar overlapping could not be detected. On the other hand, void columns and lack of filling around the bars were both revealed, even if it is probably hard to discriminate between the two situations. Classification in the frequency domain, e.g., by comparing the amplitude spectra of the radar data, rather than in the time domain was also attempted. The frequency domain signatures are depicted in Figure 2 right to demonstrate the discrimination potential of this parameter. Generally, a combined approach where the final classification is achieved by merging all the different results obtained through time domain (e.g., trace envelope, instantaneous phase and instantaneous frequency) and frequency domain parameters (e.g., phase spectra) should be explored as a way to improve the reliability of the automatic analysis. Figure 2. Time domain radar signatures (left), frequency domain radar signatures (right). A, B and C indicate the positions of the antenna for the three measurements: in front of a proper joint with bar (A), a joint without bar (B), a joint with a defect consisting in no mortar around the bar (C) 2.2 Sonic application On both specimens, a grid of measuring points with dimensions 7rows x 6columns was fixed. The measuring points were placed both on the confining columns and on the masonry panel, where the horizontally perforated units are. On this grid different types of test were carried out, direct sonic tests as well as indirect tests, in the vertical and the horizontal directions. All the transmitting/receiving points were placed on the mortar bed joints. Some measurements were performed across the masonry units only for the confining column of specimen A. Velocity analysis on the basis of direct sonic pulse velocity was performed to assess the condition of masonry. The average velocity value obtained on the bed joints NDTCE’09, Non-Destructive Testing in Civil Engineering Nantes, France, June 30th – July 3rd, 2009 without defect was about 4400 m/s, which is comparable with good quality of concrete. It is to stress that the theoretical error due to technical limitation was around 30% in this case. The average velocity computed with the measurements carried out across the masonry units was equal to 2300 m/s. Damage in masonry concentrated on the 7th bed joint was easily identified, as well as the unfilled columns of specimen B (Figure 3), since both the condition caused a decrease of average velocity, 50% and 40% respectively, in comparison with the average velocity calculated in areas of the masonry wall without defects and damage. The vertical reinforcement overlapping, placed at the bottom part of specimens A, and defects of horizontal bed joints (no filling, no cover and partial cover, respectively defect α, β and γ Figure 3) were not detectable, as expected, since the variations of velocity were within the range of variability of the measurements. However, in general, the bars overlapping caused a decrease of velocity (Figure 3 left), likely due to the worse filling with mortar in overlapping region. In the case of indirect transmission mode, the velocity values were analyzed in space versus time diagrams. Indirect tests substantially confirmed the results of direct sonic tests and moreover were able to identify the presence of reinforcement in regular bed joints, which induces a small decrease of velocity compared with unreinforced regular bed joints. Figure 3. 3 Contour map of direct sonic test: panel A (left) and panel B (right) On-site evaluation The techniques calibrated in laboratory were subsequently applied on-site. The radar and sonic methods were applied on a two-family house under construction on the Garda lake. The house is made by adopting the second construction system (Figure 1 right). The testing wall, depicted in Figure 1 (right) under construction, is characterized by presence of three vertical reinforcement position, plastic drain, damp proof course at the first bed joint and horizontal reinforcement every other two courses. The overlapping of the vertical reinforcement starting from the foundations was about three courses high. No defect was expected in this construction or at least no defect was deliberately planned for testing purpose. Therefore it was expected to have vertical and horizontal reinforcement well embedded in mortar and horizontal bed joint well executed. 3.1 GPR evaluation High frequency GPR surveying involved the collection of several vertical and horizontal profiles, so as to intersect joints with and without reinforcements as well as reinforced vertical columns. Collected data were processed and analyzed to check the stability and the repeatability of the signatures in correspondence of joints with and without reinforcements. The typical signatures derived from the statistical analysis of all the radar datasets were extracted and compared. It was concluded that the presence or absence of the reinforcements could be identified by observing the signatures in the time window between 2 and 3.2 ns (Figure 4). The application of the automatic algorithm for discriminating between the two NDTCE’09, Non-Destructive Testing in Civil Engineering Nantes, France, June 30th – July 3rd, 2009 situations was again quite successful. On the other hand, assessment of horizontal profiles gave results of poorer quality. Unfortunately, according to the design of this masonry, a vertical hole like those that host the reinforcements is placed every half width of a block (Figure 1 right). As a consequence, the detection of the reinforced vertical columns is very much disturbed by the interfering diffractions generated by the neighboring empty columns. For this reason, the discrimination between a vertical hole with bars and without bars appears rather difficult with this masonry texture and there is the need of specific tests on laboratory specimens prepared with this block design to explore a different ND approach to the problem. Figure 4. Time domain radar signatures. The antenna was in front of a proper joint with bar (left) and a joint with no bar (right) 3.2 Sonic evaluation The application of sonic technique on real masonry wall was carried out with the aim to identify defect or characteristics of real masonry wall, according to the results of laboratory tests. A grid of measuring points with dimensions 8rows x 6columns, was fixed in order to carry out sonic tests in direct transmission mode and pseudo-direct tests on vertical profiles on non-reinforced and reinforced vertical columns. The transmitting/receiving points were fixed alternatively on the bed joints and on the units for both type of test. The direct sonic tests allowed to clearly detect the presence of drain as well as the position of the vertical reinforced columns. Pseudo-direct tests - Averages 2.5 y = 3030.4x - 0.1738 2 R = 0.9693 Distance (m) 2.0 1.5 y = 2419.8x - 0.1141 2 R = 0.94 1.0 0.5 reinforced non-reinforced 0.0 0.0000 0.0002 0.0004 0.0006 0.0008 0.0010 Time (s) Figure 5. On-site tests: results of direct (left) and pseudo-direct sonic tests (right) As Figure 5 (left) shows, the drain is identified by a decrease of average velocity (about 30%), compared to the rest of masonry velocity (about 2500 m/s). The vertical reinforced column were individuated by an increase of velocity of about 37%, compared to the rest of masonry. The measurements carried out on the masonry units magnify the effect of the reinforced vertical column, whereas the measurements carried out along the bed joints NDTCE’09, Non-Destructive Testing in Civil Engineering Nantes, France, June 30th – July 3rd, 2009 highlight the effect of the presence of the drain, coherently with the typology of the tested masonry. In the case of pseudo-direct test, the average velocity values inferred from the diagrams from the measurements along the non-reinforced and the reinforced vertical columns are respectively equal to 2427 m/s and 3033 m/s (Figure 5 right). The increase of velocity along the reinforced vertical column allows detecting this type of property, with values of velocity that are consistent, in the two cases, with those found on the general direct tests (differences of -4% and -11% if the values obtained by means of the pseudo-direct tests are compared to those of the direct tests for equivalent conditions). 4 Conclusions Laboratory tests proved that high frequency GPR is a promising technique for identifying the reinforced masonry wall geometry and detecting the presence of construction defects. Specifically, GPR can detect horizontal bed joints with and without reinforcements as well as reinforcements with no mortar cover. As far as vertical columns are concerned, GPR can discriminate between a reinforced vertical column with bars and proper filling, and a defect situation like an empty column or a column with bars but no filling. The application to a real case study validated the methodology for what concerns the analysis of the horizontal reinforcements while pointed out a problem that needs further laboratory testing regarding the classification of vertical columns in some masonry design. The sonic pulse velocity analysis allows detecting general conditions of masonry but not small defects, as proved by calibration and following on-site validation carried out. Sonic pulse velocity technique can be considered as useful complementary tool for on-site assessment of: damaged portions of masonry, empty vertical columns or portions containing vertically reinforced columns. The presence of defects in the bed joints, the presence of the horizontal reinforcement or overlapping cannot be clearly detected by means of sonic tests, even if some of these conditions, such as the overlapping, need to be further investigated. At this stage, the authors believe that this methodology needs further calibration, before being successfully applied on site to real reinforced masonry walls. Acknowledgements The authors are grateful to IDS S.p.A. that supplied the Aladdin equipment for georadar experiments. The tests were carried out in the framework of DISWall: COOP-CT-2005018120: ‘Developing Innovative Systems for Reinforced Masonry Walls’. The partners of the research projects involved in the production of the described masonry system are Laterizi Alan Metauro s.r.l., Cisedil s.r.l. and Tassullo S.p.A. (Italy), Bekaert SA/NV (Belgium). References 1. Tomaževič, M. (1999) "Earthquake-resistance Design of Masonry Buildings", Imperial College Press. 2. Chang, P.C., Chiliu, S. (2003) "Recent research in non-destructive evaluation of civil infrastructures", ASCE J. Mater Civil Eng, Vol. 15, Nr 3, 2003, pp.298-304. 3. Binda, L., Saisi, A. (2001) "State of the art of research on historic structures in Italy ", Dept. pf structural engineering, Politecnico of Milan, Italy, 2001. 4. Zanzi L., Arosio D., Munda S., da Porto F., Mosele F. (2008) "Quality assessment of reinforced masonry walls by means of non-destructive tests", 2nd Canadian Conference on effective design of structures, McMaster University, Hamilton, Ontario, Canada, May 2023, (on Cd-rom).