@prefix vivo: . @prefix edm: . @prefix ns0: . @prefix dcterms: . @prefix dc: . @prefix skos: . vivo:departmentOrSchool "Applied Science, Faculty of"@en, "Civil Engineering, Department of"@en ; edm:dataProvider "DSpace"@en ; ns0:degreeCampus "UBCV"@en ; dcterms:creator "Zadeh, Aidin Abdolrahim"@en ; dcterms:issued "2009-08-07T14:17:18Z"@en, "2009"@en ; vivo:relatedDegree "Master of Applied Science - MASc"@en ; ns0:degreeGrantor "University of British Columbia"@en ; dcterms:description """Fiber reinforced polymer (FRP) repairs on concrete structures have become popular, but their long-term durability remains in question. The objective of this research project is to investigate the efficiency, practicality, and capability of non-destructive testing in assessing the durability of FRP repairs applied on concrete bridges that have been in service for some years. Three bridges were chosen as the best available case studies: (1) the Safe Bridge with sprayed FRP in British Columbia, (2) the St-Étienne-de-Bolton Bridge with four different FRP wrap products in Quebec, and (3) the Leslie Street Bridge with wrap carbon fiber reinforced polymer in Ontario. These structures represent a wide variation in severity of environmental exposure, length of time in service, type of strengthening, and type of FRP product. They also encompass flexural and shear strengthening of beams and strengthening of columns. Two non-destructive evaluation techniques, active infrared thermography and impact-echo, were chosen as the most suitable methods of bond evaluation. The results of these non-destructive techniques were compared with the actual bond strength for accuracy and consistency. To measure the actual integrity of the bond, direct mechanical pull-off tests were performed as a complementary semi-destructive method. Initial indications are that infrared thermography enables investigation of vast areas in a reasonably time-efficient manner and provides reliable qualitative results. Impact-echo is found to be ineffective in evaluating the FRP bond, as the method is incapable of assaying the material close to the impact surface. The direct mechanical pull-off test was found to be relatively accurate in providing quantitative results in environmentally mild testing conditions. Use of the direct mechanical pull-off test, in a small sample, as substantiation of the infrared thermography, significantly increases the reliability of the non-destructive technique. In general, infrared thermography showed that the bond between FRP and concrete in these structures is reasonably good in most areas, but it is strongly influenced by the placement technique, workmanship, exposure severity, and detailing."""@en ; edm:aggregatedCHO "https://circle.library.ubc.ca/rest/handle/2429/11980?expand=metadata"@en ; dcterms:extent "6149849 bytes"@en ; dc:format "application/pdf"@en ; skos:note "PRATICALITY AND CAPABILITY OF NDT IN EVALUATING LONG-TERM DURABILITY OF FRP REPAIRS ON CONCRETE BRIDGES by AIDIN ABDOLRAHIM ZADEH B.A.SC., University of British Columbia, Vancouver, Canada, 2007 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF APPLIED SCIENCE in THE FACULTY OF GRADUATE STUDIES (CIVIL ENGINEERING) THE UNIVERSITY OF BRITISH COLUMBIA (VANCOUVER) August 2009 © Aidin Abdolrahim Zadeh, 2009 ii ABSTRACT Fiber reinforced polymer (FRP) repairs on concrete structures have become popular, but their long-term durability remains in question. The objective of this research project is to investigate the efficiency, practicality, and capability of non-destructive testing in assessing the durability of FRP repairs applied on concrete bridges that have been in service for some years. Three bridges were chosen as the best available case studies: (1) the Safe Bridge with sprayed FRP in British Columbia, (2) the St-Étienne-de-Bolton Bridge with four different FRP wrap products in Quebec, and (3) the Leslie Street Bridge with wrap carbon fiber reinforced polymer in Ontario. These structures represent a wide variation in severity of environmental exposure, length of time in service, type of strengthening, and type of FRP product. They also encompass flexural and shear strengthening of beams and strengthening of columns. Two non-destructive evaluation techniques, active infrared thermography and impact-echo, were chosen as the most suitable methods of bond evaluation. The results of these non-destructive techniques were compared with the actual bond strength for accuracy and consistency. To measure the actual integrity of the bond, direct mechanical pull-off tests were performed as a complementary semi-destructive method. Initial indications are that infrared thermography enables investigation of vast areas in a reasonably time-efficient manner and provides reliable qualitative results. Impact-echo is found to be ineffective in evaluating the FRP bond, as the method is incapable of assaying the material close to the impact surface. The direct mechanical pull-off test was found to be relatively accurate in providing quantitative results in environmentally mild testing conditions. Use of the direct mechanical pull-off test, in a small sample, as substantiation of the infrared thermography, significantly increases the reliability of the non-destructive technique. In general, infrared thermography showed that the bond between FRP and concrete in these structures is reasonably good in most areas, but it is strongly influenced by the placement technique, workmanship, exposure severity, and detailing. iii TABLE OF CONTENTS ABSTRACT........................................................................................................................ ii TABLE OF CONTENTS................................................................................................... iii LIST OF TABLES............................................................................................................ vii LIST OF FIGURES ......................................................................................................... viii 1 INTRODUCTION ...................................................................................................... 1 1.1 Introduction................................................................................................. 1 1.2 Background of non-destructive evaluation ................................................. 2 1.3 Stress wave methods ................................................................................... 3 1.3.1 Ultrasonic pulse velocity............................................................................... 3 1.3.2 Acoustic emissions method........................................................................... 3 1.3.3 Short pulse radar ........................................................................................... 4 1.3.4 Impact-echo method...................................................................................... 4 1.4 Radioactive/nuclear methods ...................................................................... 5 1.4.1 Radiography and radiometry......................................................................... 5 1.4.2 Neutron–gamma method............................................................................... 6 1.4.3 Infrared thermography .................................................................................. 6 2 INFRARED THERMOGRPHY ................................................................................. 7 2.1 Fundamental theory .................................................................................... 7 2.1.1 Passive thermography ................................................................................. 10 2.1.2 Active thermography .................................................................................. 11 2.2 Infrared thermography technique.............................................................. 11 2.2.1 Externally applied thermal field thermography .......................................... 11 2.2.1.1 Lock-in thermography ......................................................................... 12 2.1.1.2 Pulse thermography .................................................................................. 13 2.2.2 Externally applied vibration field (vibrothermography)............................. 15 2.3 Automated signal analysis systems........................................................... 15 2.3.1 Neural networks for defect detection .......................................................... 16 2.3.2 Region growing algorithms......................................................................... 16 iv 2.4 Trial laboratory tests ................................................................................. 16 2.4.1 Experimental procedure .............................................................................. 17 2.4.2 Results and conclusions .............................................................................. 18 3 IMPACT-ECHO ....................................................................................................... 20 3.1 Fundamental theory .................................................................................. 20 3.2 Signal analysis .......................................................................................... 23 3.3 Impact-echo transducer ............................................................................. 27 3.4 Wave speed ............................................................................................... 28 3.4.1 Direct method of P-wave speed measurement............................................ 29 3.4.2 P-wave speed measurement from impact-echo testing on plates with known dimensions ...................................................................................... 31 3.5 Transformation from time to frequency domain....................................... 31 3.6 Interfacial wave reflections and refractions .............................................. 32 3.6.1 Solid/air interfaces ...................................................................................... 33 3.6.2 Solid/solid interfaces................................................................................... 33 3.6.3 Diffractions ................................................................................................. 35 3.7 Spring force effects ................................................................................... 35 3.8 Choosing the proper impactor................................................................... 36 4 SITES OF INTEREST AND METHODOLOGY .................................................... 38 4.1 Sites of interest.......................................................................................... 38 4.1.1 Safe Bridge (Vancouver Island, British Columbia) .................................... 38 4.1.1.1 Description and background................................................................ 38 4.1.1.2 PCCB bridges’ design deficiencies ..................................................... 39 4.1.1.3 Retrofitting........................................................................................... 39 4.1.1.4 Climatic conditions.............................................................................. 42 4.1.2 St-Étienne Bridge (St-Étienne-de-Bolton, Québec).................................... 43 4.1.2.1 Description and background................................................................ 43 4.1.2.2 Climate conditions............................................................................... 44 4.1.3 Leslie Street Bridge (Toronto, Ontario)...................................................... 46 4.1.3.1 Description and background................................................................ 46 v 4.1.3.2 Climatic conditions.............................................................................. 49 4.2 Methodology and experimental program.................................................. 50 4.2.1 Experimental program ................................................................................ 50 5 TEST RESULTS AND DISCUSSION .................................................................... 56 5.1 Non-destructive tests................................................................................. 56 5.1.1 Infrared thermography ................................................................................ 56 5.1.1.1 Safe Bridge, British Columbia............................................................. 56 5.1.1.2 St-Étienne-de-Bolton Bridge, Québec ................................................. 59 5.1.2 Impact-echo................................................................................................. 60 5.2 Semi-destructive mechanical pull-off test ................................................ 63 5.2.1 Pull-off results of flexural members (Safe Bridge)..................................... 65 5.2.2 Pull-off results from vertical weight support members .............................. 70 5.3 Possible causes of FRP–concrete bond deterioration ............................... 74 5.3.1 Insufficient surface preparation .................................................................. 75 5.3.2 Water seepage ............................................................................................. 75 5.3.3 Heat of coring ............................................................................................. 78 5.3.4 Snow clearing process................................................................................. 79 6 CONCLUSIONS AND FUTURE RESEACH RECOMMENDATIONS ............... 80 6.1 Conclusions............................................................................................... 80 6.1.1 Infrared thermography ................................................................................ 80 6.1.2 Impact-echo test .......................................................................................... 81 6.1.3 Mechanical pull-off test .............................................................................. 82 6.2 Future research and recommendations...................................................... 83 6.2.1 Infrared thermography ................................................................................ 83 6.2.2 Impact-echo technique ................................................................................ 84 6.2.3 Direct mechanical pull-off test.................................................................... 84 REFERENCES ................................................................................................................. 86 APPENDIX A. FRP APPLICATION PROCEDURE (TYFO SYSTEMS)..................... 90 APPENDIX B. INFRARED THERMOGRAPHY RESULTS ........................................ 91 B.1 Safe Bridge............................................................................................................. 91 vi B-2 St-Étienne Bridge................................................................................................... 96 APPENDIX C. IMPACT_ECHO TEST RESULTS ........................................................ 98 C.1 St-Etienne Bridge ................................................................................................... 98 C.2 Leslie Street Bridge.............................................................................................. 104 vii LIST OF TABLES Table 2.1 Typical thermal review of infrared thermography……………………………..9 Table 5.1 Tensile bond strenth from all three bridges ...................................................... 64 Table 5.2. Pull-off test results from column 1 of the St-Etienne Bridge .......................... 66 Table 5.3. Pull-off test results from column 2 of the St-Etienne Bridge .......................... 66 Table 5.4. Pull-off test results from column 3 of the St-Etienne Bridge .......................... 67 Table 5.5. Pull-off test results from column 4 of the St-Etienne Bridge .......................... 67 Table 5.6. Pull-off test results from the Safe Bridge mid-span ........................................ 68 Table 5.7. Pull-off test results from the Safe Bridge eastern near support end span ........ 69 Table 5.8. Pull-off test results from the Safe Bridge western near support end span ....... 69 Table 5.9. Pull-off test results from column 1 of the Leslie Street Bridge ....................... 73 Table 5.10. Pull-off test results from column 1 of the Leslie Street Bridge ..................... 74 viii LIST OF FIGURES Figure 2.1. Heat diffusion into the material. Note how the debonded spot acts as an insulator............................................................................................................. 8 Figure 2.2. A typical setup for modulated thermography................................................. 13 Figure 2.3. A typical setup for pulse thermography ......................................................... 14 Figure 2.4 FRP reinforced slab used in trial laboratory testing ........................................ 17 Figure 2.5 Left: Thermographic image of trial slab; Right: Teflon sheet arrangement underneath FRP............................................................................................... 18 Figure 3.1. Schematic representation of P-, S-, and R-wave fronts generated by an elastic impact.............................................................................................................. 21 Figure 3.2. Impact-echo testing setup. Note that as the impact is generated by the impactor, surface displacements are picked up by the transducer .................. 23 Figure 3.3. Schematic showing the movement of stress waves shortly after impact at time t, with corresponding ideal time vs. voltage graph. Notice the P-wave passing by the transducer ............................................................................................. 24 Figure 3.4. Schematic showing the movement of stress waves at time t + e, with the corresponding ideal time vs. voltage graph. Notice the R-wave passing by the transducer ........................................................................................................ 25 Figure 3.5. Schematic showing the movement of stress waves at time t + 2e, with the corresponding ideal time vs. voltage graph. Notice the reflected P-wave passing by the transducer ................................................................................ 26 Figure 3.6. Time vs. voltage graph, showing the periodic movement of P-wave in the specimen ......................................................................................................... 27 Figure 3.7. A dual transducer, necessary for measuring P-wave speed............................ 29 Figure 3.8. Simplification of P-wave movement in plate-like structures, using the ray light analogy.................................................................................................... 33 Figure 4.1. Safe Bridge, Vancouver Island, BC ............................................................... 38 Figure 4.2. Section and plan view of the Safe Bridge showing the girder numbering and geometric dimensions ..................................................................................... 41 ix Figure 4.3. Sectional view of a box girder from the Safe Bridge showing naming of the legs .................................................................................................................. 41 Figure 4.4. Safe Bridge's annual temprature profile ......................................................... 42 Figure 4.5. Safe Bridge's annual snowfall profile............................................................. 43 Figure 4.6. Plan view of the St-Étienne Bridge showing the numbering of the columns (note that the figure is not to scale)................................................................. 44 Figure 4.7. Regular usage of de-icing salts induce a significant amount of corrosion on the St-Étienne Bridge...................................................................................... 45 Figure 4.8. St-Étienne Bridge's average annual temperature profile ................................ 45 Figure 4.9. St-Étienne Bridge's average snowfall profile ................................................. 46 Figure 4. 7. Leslie Street Bridge. Note the level of corrosion in steel flexural members ......................................................................................................................... 47 Figure 4.8. Leslie Street before and after the application of FRP repairs......................... 48 Figure 4.9. Leslie Street Bridge. Columns on which the direct pull-off test was performed are marked....................................................................................................... 48 Figure 4.10. Toronto's annual average temperature profile .............................................. 49 Figure 4.11. Average annual snowfall profile in Toronto ................................................ 50 Figure 4.12. Schematic of testing apparatus ..................................................................... 51 Figure 4.13. Gridding and random selection of core locations......................................... 52 Figure 4.14. The position of the column grids with respect to the road ........................... 52 Figure 4.15. Heating of FRP surface as part of active thermography............................... 53 Figure 4.16. Securing test disks during epoxy setting ...................................................... 54 Figure 4.17. Pull-off testing using a Swiss-made DVNA Haftprufer Z16 ....................... 55 Figure 4.18 The coring setup ............................................................................................ 55 Figure 5.1. (a) Growth of fungi detected by thermography. (b) A circular and rectangular demination found on the side of the girder. (c) Detection of water pathway and fiber optics by thermography. Note the location of a pull-off test which indicated zero bond strength. (d) The water seepage pathway clearly visible in the thermographic image................................................................................. 58 x Figure 5.2. Thermographic image of column 1 (shown on right) with bond strength values acquired by the direct pull-off test....................................................... 60 Figure 5.3. Doubling of stress on the bond due to presence of debonded areas ............... 60 Figure 5.4. An ambiguous impact-echo signal spectrum from location J0 on column 2 of the Leslie Street Bridge. The graph does not provide any significant information on bond condition nor in situ material discontinuity .................. 62 Figure 5.5. An impact-echo signal spectrum from location J6 on column 2 of the Leslie Street Bridge. The graph clearly shows the presence of in situ steel reinfocement at 43.5 cm below the impact surface......................................... 63 Figure 5.6. Modes of failure in a direct mechanical pull-off test: (a) failure of concrete substrate, (b) failure of FRP, (c) failure of substrate mortar repair, (d) combined failure of substrate concrete and FRP ............................................ 72 Figure 5.7. The presumed pathway of water seepage under the sprayed FRP applied on the Safe Bridge................................................................................................ 76 Figure 5.8. EDX and X-ray scans of the FRP underside from core 5A-east, showing presence of rust particles and salt crystals ...................................................... 77 Figure B.1.1 1A-East, Safe Bridge ................................................................................... 91 Figure B.1.2 Girder 2A-East, Safe Bridge ........................................................................ 91 Figure B.1.2 1B-East, Safe Bridge.................................................................................... 91 Figure B.1.4 Girder 2B-East, Safe Bridge ........................................................................ 91 Figure B.1.3 Girder 5A-East, Safe Bridge ........................................................................ 92 Figure B.1.4 Girder 5B-East, Safe Bridge ........................................................................ 92 Figure B.1.5 Girder 8A-East, Safe Bridge ........................................................................ 92 Figure B.1.6 Girder 5B-East, Safe Bridge ........................................................................ 92 Figure B.1.7 Girder 8A-East, Safe Bridge ........................................................................ 92 Figure B.1.8 Girder 8B-East, Safe Bridge ........................................................................ 92 Figure B.1.9 Girder 10A-East, Safe Bridge ...................................................................... 93 Figure B.1.10 Girder 10B-East, Safe Bridge .................................................................... 93 Figure B.1.11 Girder 1A-West, Safe Bridge..................................................................... 93 Figure B.1.12 Girder 1B-West, Safe Bridge..................................................................... 93 xi Figure B.1.13 Girder 2A-West, Safe Bridge..................................................................... 93 Figure B.1.14 Girder 2B-West, Safe Bridge..................................................................... 93 Figure B.1.15 Girder 5A-West, Safe Bridge..................................................................... 94 Figure B.1.18 Girder 5B-West, Safe Bridge..................................................................... 94 Figure B.1.16 Girder 8A-West, Safe Bridge..................................................................... 94 Figure B.1.17 Girder 10A-West, Safe Bridge................................................................... 94 Figure B.1.18 Girder 10B-West, Safe Bridge................................................................... 94 Figure B.1.19 Girder 1A-Middle, Safe Bridge ................................................................. 94 Figure B.1.20 Girder 1B-Middle, Safe Bridge ................................................................. 95 Figure B.1.21 Girder 2A-Middle, Safe Bridge ................................................................. 95 Figure B.1.22 Girder 2B-Middle, Safe Bridge ................................................................. 95 Figure B.2.1 Column1-Exposed, St-Etienne……………………………………………..96 Figure B.2.2 Column1- Inner, St-Etienne..……………………………………………....96 Figure B.2.3 Column1- Exposed, St-Etienne.………………………………………........96 Figure B.2.4 Column1- Inner, St-Etienne.……………………………………………….96 Figure B.2.5 Column1- Exposed, St-Etienne…………………………………………….96 Figure B.2.6 Column1- Inner, St-Etienne………………………………………………..96 Figure B.2.7 Column1- Exposed, St-Etienne…………………………………………….97 Figure B.2.8 Column1- Inner, St-Etienne………………………………………………..97 Figure C.1. Column #1- D16 ............................................................................................ 98 Figure C.2. Column #1- G17 ............................................................................................ 98 Figure C.3. Column #1- D13 ............................................................................................ 99 Figure C.4. Column #1- F12 ............................................................................................. 99 Figure C.5. Column #1- B10............................................................................................. 99 Figure C.6. Column #1- B1............................................................................................. 100 Figure C.7. Column #1-F3 .............................................................................................. 100 Figure C.8. Column #1-A4 ............................................................................................. 100 Figure C.9. Column #1- H6 ............................................................................................ 101 Figure C.10. Column #1- F7 ........................................................................................... 101 Figure C.11. Column #3- A11 ........................................................................................ 101 xii Figure C.12. Column #-E10............................................................................................ 102 Figure C.13. Column #3- G11 ........................................................................................ 102 Figure C.14. Column #3- E15......................................................................................... 102 Figure C.15. Column #3-H16 ......................................................................................... 103 Figure C.16. Column #3- D17 ........................................................................................ 103 Figure C.17. Column #3- D6 .......................................................................................... 103 Figure C.18. Column #3- A6 .......................................................................................... 104 Figure C.19. Column #3- F2 ........................................................................................... 104 Figure C.20. Column #2- D9 .......................................................................................... 104 Figure C.21. Column #2- J9............................................................................................ 105 Figure C.22. Column #2- J6............................................................................................ 105 Figure C.23. Column #2- J10.......................................................................................... 105 Figure C.24. Column #2- J0............................................................................................ 106 Figure C.25. Column #2- C0........................................................................................... 106 Figure C.26. Column #2- A2 .......................................................................................... 106 Figure C.27. Column #2- C3........................................................................................... 107 Figure C.28. Column #2- J5............................................................................................ 107 1 1 INTRODUCTION 1.1 Introduction The majority of modern Canadian transportation infrastructure was built during the 1960s, and many of these structures are now showing signs of deterioration and distress. The situation in many other countries is equally dire, and many structures need replacement. To delay bridge replacement as long as possible, application of external reinforcing and strengthening elements such as fiber reinforced polymers (FRPs) has become popular in Canada and indeed around the world. FRP is generally applied locally as a patch to enhance strength. Primary testing of the material under short-term static or static-cyclic loading indicated an increase in the flexural and compressive capacity of structural elements [1]. Beams are generally reinforced in shear and flexure, and columns are wrapped to increase strength through confinement and to increase structural durability. A variety of FRP products with varying physical and mechanical properties have been introduced to the market [2]. As an added advantage FRPs provided an insulating layer against corrosion in cases of direct exposure to salts. Unlike concrete and reinforcing steel, little is known of the long-term durability of FRPs and the effects of the environment on these materials. A comprehensive study on the durability of glass fiber reinforced polymer (GFRP) rods used as internal reinforcement in in-service bridges was carried out by Mufti et al. in 2006 [3]. It was concluded that GFRPs used in bridges had shown no signs of deterioration and alkali attack had not occurred. Although this research implied that GFRP is sufficiently durable, but the scope of this research did not include the durability study of the externally applied FRP material on concrete. Thus, the durability of FRPs used as externally bonded strengthening coats remains poorly understood. One concern is their durability when exposed to ultraviolet light and heat [4], but proper coatings have been developed to minimize damage from such exposure. Another concern is durability of the bond between concrete and the FRP. 2 To investigate this, direct mechanical pull-off tests were carried out on four bridges in Canada [5]. The results from this study suggested that the FRP–concrete bond is durable in most environmental conditions, but factors such as workmanship, properties of cement substrate, and detailing appear to significantly affect the bond durability. The aim of this study is to investigate the efficiency, practicality, and capability of non- destructive testing methods in assessing the strength of the bond between FRPs and the bridge concrete to which they are applied. This was done by using the results from the well-established semi-destructive direct mechanical pull-off test as a quantitative bond measure for comparison, 1.2 Background of non-destructive evaluation Non-destructive testing was defined for the first time in the 1970s as “examination of an object or material that did not render it unfit for use”. Terms used with that early definition were non-destructive testing (NDT), non-destructive inspection (NDI), and, non-destructive examination (NDEx) [6]. The presence of flaws or discontinuities in a material was determined by looking at or through an object. The term NDE (non- destructive evaluation) emerged in the 1980s and included the terms NDT and NDI described earlier. However, the definition was later expanded in the field of metallurgy to include the classification of a discontinuity by its size, shape, type, and location. With advancements in fracture mechanics, NDE using highly sensitive equipment can now provide a non-destructive characterization of metallic microstructure as well. There are many non-destructive testing techniques, each using a different mechanism to relay material conditions. However, capabilities and the level of practicality in all these methods are different. NDT is mainly divided into three subcategories: (1) stress wave methods, (2) nuclear methods, and (3) magnetic and electric methods. In this chapter, an overview of stress wave methods and nuclear methods, which are the non-destructive evaluation methods most widely used on concrete structures, are briefly discussed. Two non-destructive methods with the most promising potential of success in evaluating FRP 3 bond conditions, chosen for experimentation, are comprehensively presented in Chapters 2 and 3. 1.3 Stress wave methods Stress wave methods introduce a type of wave, either mechanical or sound waves or electromagnetic waves. The waveforms are then recorded by one or a series of receivers, and depending on a specific wave behavior, valuable information about the material of interest can be inferred from the data. Most of the stress wave methods are capable of providing subsurface images of the specimen, allowing the users to visually assess flaws and faults within the material. The most popular of the stress wave methods are described below. 1.3.1 Ultrasonic pulse velocity The ultrasonic pulse velocity (UPV) technique evaluates material integrity by measuring the propagation velocity of an ultrasonic high-frequency compression P-wave through a solid. The wave velocity is calculated using the time taken by the pulse to travel the measured distance between the transmitter and the receiver [7]. This velocity generally varies through the material, depending on the density and elastic properties of the material. As material quality is related to elastic properties, UPV is an excellent method to use to determine the extent of decay in a given specimen [8]. 1.3.2 Acoustic emissions method Acoustic emission (AE) monitoring of concrete structures provides qualitative information about active microcracking. An acoustic emission is a localized rapid release of strain energy in a stressed material. 4 The development of cracks and microcracks inside a solid specimen generates a sudden release of strain energy and is accompanied by the emission of audible noise. With the help of AE sensors or microphones, amplifiers, and a data acquisition unit, audible sounds at stress levels of 10% of the ultimate strength are detectable [7]. Recording the acoustic emissions from within the material allows approximation of the location of cracks by triangulation and overall structural quality control [9]. 1.3.3 Short pulse radar Short pulse radar (SPR) is a powerful and rapid technique that uses electromagnetic (EM) waves as opposed to ultrasonic pulses. It is often employed in the oil industry for probing solids such as rock or soil. It is also extremely useful in detection of water and water saturated material. SPR operates, on principle, by taking advantage of the differences in reflection and scattering properties of EM waves to create two-dimensional images of subsurface sections of a specimen [10]. 1.3.4 Impact-echo method Impact-echo is a non-destructive method of material testing. By recording the travel time of impact generated transient stress waves using a sensitive broadband transducer, it provides valuable information about continuity and integrity of subsurface material. As the impact pulse propagates through the material, it is partially reflected when it reaches another interface, such as a void space or a crack. When the reflected wave arrives back at the concrete surface, it causes a surface displacement. Measurements of this surface displacement then provide information about the internal state of the material. 5 Impact-echo is often used for a preliminary survey of critical areas for locating anomalies. Imaging of these anomalies may then be performed using more comprehensive testing methods. The method requires a relatively high level of expertise and training compared to other NDT methods. 1.4 Radioactive/nuclear methods These methods are based on introducing subatomic particle waves into a material and then collecting the subatomic particle waves after they have passed through the material. It is widely used in applications such as medical imaging and criminal forensics. However, owing to the health risks associated with exposure to radioactive material and sensitivity of the human tissue, these techniques are primarily limited to laboratory and controlled environmental conditions. 1.4.1 Radiography and radiometry Radiography utilizes the upper components of the electromagnetic spectrum, namely, X- rays and gamma rays, to examine the internal conditions of material. In principle, the technique captures shadows of the material being inspected when X-rays are passed through it. Since the radiation penetrating the material is attenuated to varying degrees by the constituents of the internal structure, the shadow contains details within. Use of X- rays and computer tomography (CT) could generate three-dimensional models of material. Radiometry is similar to radiography in principle, as the method takes advantage of the attenuation of gamma rays generated by a radioactive isotope. However, rather than using an electron sensitive film, which visually illustrates disparities in density within the concrete, radiometry uses a gauge such as a Geiger counter to determine the intensity of emerging radiation [11]. 6 1.4.2 Neutron–gamma method The neutron–gamma method is the most underdeveloped of the radiation/nuclear methods applied to concrete testing. Consequently, at this point, it is almost exclusively limited to laboratory testing of specimens. The method requires a neutron source along with a gamma ray collection and counting system. The neutron–gamma method irradiates material with neutrons. The interaction of neutrons with various elements within the material induces emissions of gamma rays having energy characteristics of that specific element. The number of gamma rays of a specific energy detected over a certain period of time can then be related to the amount of a specific element present in a sample [12]. 1.4.3 Infrared thermography Infrared thermography (IT) is a safe non-contact non-destructive material evaluation method where isothermal contours of steady or transient thermal effects are constructed from measurements of the material’s infrared energy or heat emissions. IT technique is based on the theory that the flow of heat through a material is disturbed in the presence of internal discontinuities, cracks, delaminations, or other anomalies. Such anomalies result in local concentration of heat and hence creation of an area where relative temperature variation exists; these spots, commonly known as” hot spots”, are easily detectable by infrared sensors. The IT technique is described in detail in chapter 2. 7 2 INFRARED THERMOGRPHY Infrared thermography (IT) is a non-contact non-destructive material evaluation method where isothermal contours of steady or transient thermal effects are constructed from measurements of a material’s infrared energy or heat emissions. IT technique is based on the theory that the flow of heat through a material is disturbed in the presence of internal discontinuities, cracks, delaminations, or other anomalies. Such anomalies result in local concentration of heat ,commonly known as “hot spots”, and hence creation of an area where relative temperature variation exists, which is easily detectable by infrared sensors. Initially implemented and used as an NDT technique in 1950 [14], IT is generally regarded as a qualitative technique of inspection. However, after development of cost- effective and high-resolution imaging sensors, a significant amount of research has been carried out to develop systems that automatically analyze IT signals using computer algorithms [20]. 2.1 Fundamental theory The principle of IT is to heat up one surface of the spatial region of interest and as heat diffuses into the material, the presence of discontinuities and flaws within the mass act as heat insulators and cause interruptions in the heat flow, thus leading to generation of hot spots. According to ASTM Standard D4788-03 [13], a defect is identified as areas with a temperature variation of 0.5 °C relative to their surroundings. Figure 2.1 provides a schematic of the heat flow process shortly after the heating process has commenced. 8 Figure 2.1 Heat diffusion into the material. Note how the debonded spot acts as an insulator. Conduction of heat within any medium or between two materials is described by the Fourier’s law: At Tk Q × ∆× = (Equation 2.1) where Q is thermal energy, k is the material’s thermal conductivity, A is the surface area on which the energy is incident, t is the specimen thickness, and ∆T is the temperature gradient. Fourier’s law indicates that materials with a high k-value, such as metals, can conduct heat quicker than an air gap or even a vacuum, which is characterized by a near zero k- value. FRP composite materials have much lower thermal conductivity as compared to metals but a relatively higher k-value with respect to air. 9 Table 2. 1 Typical thermal conductivities of materials Material Thermoal conductivity (W/m K) Aluminum 175-2.50 Mild Steel 40-75 Concrete 0.8-1.2 Brick 0.5-1.6 Typical polymers used in FRP 0.25-0.4 Standard/intermediate modulus PAN carbon fibres 20 High modulus PAN carbon fibres 50-80 UHM pitch carbon fibres 400-1100 GFRP (Vf = 60%) 0.25-0.32 CFRP (Vf = 60%) 0.8-1.4 As the rate of heat diffusion is a function of the properties of material and material flaws such as cracks and delaminations can be considered as air gaps with a low k-value within a solid medium, IT can reveal near surface flaws through detection of discontinuities in the temporal profile of the surface temperature. In general, the three-dimensional thermal response of any semi-infinite solid can be defined with Equation 2.2. δt δT δz Tδ δy Tδ δx Tδ α =++ 2 2 2 2 2 2 (Equation 2.2) where α is the thermal diffusivity of the material. Equation 2.2 can be further simplified by cancelling the lateral heat flow components once the point of interest in the solid is safely away from the edges and the solid does not contain any flaws. However, the presence of any near-surface anomalies in the direction of the heat wave travel would result in entrapment of incident thermal energy between the anomaly and the solid surface. Such phenomena diverges the heat wave propagation laterally outward toward the relatively cooler areas around the defect. Under these circumstances, the three-dimensional diffusion equation (Equation 2.2) can be expressed as: 10 δt δT δy Tδ δx Tδ α =+ 2 2 2 2 (Equation 2.3) It must be noted that Equations 2.2 and 2.3 are only idealized description of the heat flow pattern. Consequently, realistic differences in thermal diffusions rates within a material, in the presence of any abnormalities, cause non-uniformities in the temperature profiles, producing a defect signature detectable by IT technique. Thereby, it is inevitable that IT is constrained by thermal response of material and without thermal deviations no detection is possible. Hot spots are detected by using infrared cameras. These infrared cameras do not directly measure temperature. Instead, they operate by measuring the amount of infrared radiation from a material body and converting that value to temperature using the emissivity of that specific specimen. Emissivity value is defined as the ratio the energy radiated by a body at a given temperature to the energy radiated by a black body at the same temperature. Emissivity has a maximum value of 1, which is never actually allocated to anything in nature, as it corresponds to a perfect black body (a mass that absorbs all heat radiation) and a minimum value of 0, referred to a perfect deflector of heat radiation. Two general modes of thermography are commonly used, i.e., active and passive. 2.1.1 Passive thermography In passive thermography the discontinuities in temperature profiles are detected by the heat inputs from ambient environmental sources such as sunlight, heat generated from local sources, or frictional heat from moving parts in machinery. In the case of bridges, the sunlight is generally the sole source of heat. This reduces the amount of equipment required for thermography, such as heating instruments and their necessary accessories such as generators, cables, etc., resulting in significant time and cost savings. 11 2.1.2 Active thermography Active thermography, in which an external heat source is used to introduce energy to the specimen, has a valuable advantage over the passive method. In this method, the duration, magnitude, and frequency of the heat source are known entities and thus can be used to characterize defects on a more quantitative basis. Previous studies have shown that active thermography methods have yielded better results than the passive methods [15]. Most concrete structures, under normal conditions, have near perfect temperature distribution on their surfaces. Hence, detection of areas of heat concentration without introduction of heat from an external source is difficult. This issue is especially relenent in cases where structural elements are indoors or isolated from ambient heating source such as sunlight. 2.2 Infrared thermography technique IT technique has proved itself more effective in non-destructive testing of certain material systems such as FRP composites, insulation materials, and FRP composites bonded to metals, composite–concrete bonds, and flaws in concrete. As there are several limitations and constrains such as the maximum depth of heat diffusion when it comes to thermography, IT is only incorporated as a NDT in nearsurface material evaluations. Inspection of each material or material system requires some variations in the technique for optimal result, as the infrared absorption properties change with material type and geometry. Several methods are currently available for generation of thermal fields; two of the most common ones are externally applied thermal field (EATF) and stress generated thermal field (vibrothermography). 2.2.1 Externally applied thermal field thermography EATF thermography is carried out by studying the transient thermal gradient after application of external heat to the solid surface. The mode of thermography could be both 12 passive and active, depending on the external source of heat used. The presence of flaws causes perturbation in the heat flux and temperature field, forming curved isotherms and non-uniform temperatures near the flaw. EATF can be further categorized into parallel and normal modes, depending on the orientation of the material flaw. Delaminations and other subsurface discontinuities disturb the heat flux normal to the surface under inspection, leading to generation of hot spots on the heated surface, whereas presence of cracks and surface flaws produce local convexities in the isotherms by parallel alternation and disturbance of the heat flux [16]. The passive EATF method using sunlight as an ambient source of external heat is preferred for large structures, as this method has proven effective in measurement of the thermal diffusivity variation and heat loss in sections of an entire building [17]. However, for more qualitative measurements and characterization of defects, such that the magnitude and duration of heating could be controlled to better suit the system being inspected, active thermography must be employed. In EATF, heat is generally applied using a light bulb, heat guns, electric heater, quartz lamps, etc. Two of the most common types of EATF are lock-in thermography and pulse thermography. 2.2.1.1 Lock-in thermography Lock-in thermography, also known as modulated thermography, uses sinusoidal temperature simulation to heat the surface of the specimen. The magnitude and phase of the heat wave is adjustable by changing the input frequency. The method often uses either a modulated laser beam or a modulated heat beam to inspect a point or an area on the surface. Modulation of heat wave and thermal stimulation of specimen surface instigates oscillation of temperature field inside the solid; this is detectable by an infrared camera. An amplifier is used to synchronize the input and output signals, and the magnitude and phase of the I/O signals are computed with respect to the reference modulation. Recently, however, advanced imaging techniques have been used to acquire output signals through analysis of the digitized data. This analysis is carried out by calculating the signal 13 magnitude from the local optical and infrared surface properties and signal phase from wave propagation time delay. Using digitized data, subsurface flaws are detected by monitoring data abnormalities in comparison with data from an unflawed region [18]. The only downside of the lock-in thermography method is the sophistication of the setup necessary for producing sinusoidal modulated heat waves. Figure 2.2 shows a schematic for a modulated thermography setup. Figure 2.2 A typical setup for modulated thermography. 2.1.1.2 Pulse thermography Pulse thermography (PT) is by far the most common technique used for inspection of FRP composites and concrete (see Figure 2.3). Developed directly from the fundamental theory behind infrared thermography, PT detects air voids and material flaws by pointing out areas of temperature differentials. By launching either a “cold” or ”warm” simulation, PT generates a thermal front. The thermal front is usually deployed by means of lamps, laser beams, flashes, or air jets and then propagates through the sample under inspection by diffusion. The duration of the pulse depends on the thickness of material to be probed and its thermal properties. As the rate of diffusion is different between flawed and intact 14 regions and the pulse has a constant flux, application of the thermal pulse to the surface continues until a thermal asymptote detectable by infrared sensors is generated. The thermal propagation time, t, and the depth, z, of a defect are often related by α 2z t ≈ (Equation 2.4) where α is the material thermal diffusivity of material expressed by C K ρ α = (Equation 2.5) in which K is the thermal conductivity, ρ the mass density, and C is the specific heat of material. The PT setup consists of a heat source, an infrared sensor (camera), and a data acquisition unit. Figure 2.3 A typical setup for pulse thermography. 15 2.2.2 Externally applied vibration field (vibrothermography) In vibrothermography, forced steady-state mechanical vibrations are used as an external stimulus to produce thermal gradients between intact areas and those with defects. As the behavior of most material in the plastic region of the load range is characterized by hysteresis loops, the transmitted vibratory mechanical energy is converted to heat energy through non-conservative micro-mechanical deformation processes such as dislocation motion, impurity diffusion, and other complex atomic activity. Thermal patterns generated from this mechanical energy are altered around defects, leading into generation of hot spots. Hot spots are generated because cracks, flaws, delaminations, and other flaws in the material act as stress concentrators under applied loading/unloading cycles. Therefore, cyclic loading allows for generation of enough heat to facilitate detection of defects and flaws within a material. In vibrothermography, the amplitude and frequency of the vibration generated by the piezoelectric shaker plays a significant role. High frequency of oscillation is generally used to enhance the efficiency of the heating, whereas a low-frequency amplitude modulation tends to improve depth range [19]. 2.3 Automated signal analysis systems For quantitative flaw detection and direct interpretation of the condition, depth, size, and progression of the defects, automated IT techniques have been developed through which the temperature decay curve of the specimen surface is continuously monitored and recorded by one or multiple external recording devices. Automated signal analysis systems (ASASs) either analyze IT signals based on predefined automated data processing algorithms that provide consistency in signal analysis independent of the subjective human signal interpretation, independent of the mode of thermography (active or passive) and prior knowledge of the sample, or require identification of a defect-free reference region by the operator to be able to distinguish regions with defects. 16 Use of contrast definition is very common in automated systems where PT is the method of thermography. It must, however, be noted that although ASASs seem fairly accurate and efficient in theory and have outperformed the traditional thermography methods for provision of quantitative results in the laboratory; they are hard to implement and have not shown optimistic results in the field [20]. Two of the most commonly used algorithms for ASASs are neural networks for defect detection and region growing algorithms. 2.3.1 Neural networks for defect detection Neural networks (NN) are knowledge based systems which are engineered to imitate human knowledge. NN can be defined as computational structures built out of a series of simple process units called neurons. Neurons receive and convey information by different synaptic weights that change the transmitted signals. The network performance is independent of the number of neurons, and the network requires a learning phase prior to operation. The NN learning phase is generally carried out using a resilient back propagation training algorithm and does not require human supervision [20]. 2.3.2 Region growing algorithms Region growing algorithms process thermal images by expansion of pixels around regions marked as anomalies through thermal image segmentation. Image segmentation is accomplished by dividing an image into its constituent parts. Most segmentation algorithms are based on similarity or discontinuity features such as points, edges, and lines between different pixels in the image. They are generally categorized into four major classifications: threshold based, edge detection, region growing, and hybrid methods. Once the image is filtered and segmented, it is post-processed using a morphological program for a more clear display of defects [20]. 2.4 Trial laboratory tests To better understand and become familiar with the application of thermography prior to any field testing and also to ensure the suitability of the method as a non-destructive test 17 for FRP bond evaluation, an experimental procedure was developed in the laboratory to test the limits of thermography when detecting flaws and areas of bond deterioration in FRP repairs on concrete. The aim of such a test was to determine any variables and factors of importance when testing FRP repairs. 2.4.1 Experimental procedure To better serve the experimental goals, a FRP–concrete bond system was built with manually introduced flaws on predetermined locations to simulated areas of bond inefficiency. A 1000 mm × 500 mm × 100 mm concrete slab was cast (Figure 2.4) and repaired using two layers of Tyfo glass system wrap on one side (See Appendix A for the FRP warping procedure) after sandblasting and water-jet washing the sample surface to enhance bonding properties of the sample surface. A series of 0.5 mm thick Teflon sheets with a constant 100 mm width and varying lengths were placed between the FRP and the concrete to simulate areas of FRP delaminations. The Teflon sheets were placed as far away as possible from the sample edges to avoid any possible edge effects. Figure 2.4 FRP reinforced slab used in trial laboratory testing. Teflon sheets were chosen for their low thermal diffusivity, suitability and variability of thickness, flexibility, and their capability to be easily cut down to any dimension. Active 18 mode pulse thermography was intended as the method of thermography, and a dual 500 W commercial halogen lamp was used to generate the necessary heat pulse. In theory, heat should homogenously diffuse through FRP into concrete once the FRP surface is uniformly heated. On the other hand, owing to the low diffusivity of Teflon sheets, uniform heating of the FRP surface should result in entrapment of heat and generation of hot spots at locations where Teflon sheets are embedded. Once surface heating is halted, these areas of heat concentration should be easily detectable by an infrared camera. In this case, A FLIR-EX300 infrared camera was used to measure the surface heat emissions. 2.4.2 Results and conclusions The test was carried out several times, as it became apparent that the experimental results vary depending on the duration of the heat pulse. In cases where FRP was heated for periods of less than 45s, the infrared camera did not recognize any differentiation of infrared emission across the specimen, as the input heat was insufficient to diffuse into the concrete at any location. On the other hand, in cases where heat exposure was more than 80 s, the sample surface temperature rose so much that any heat entrapment at the FRP–concrete interface could no longer be detected by the infrared camera. It was thus concluded that the heating interval should be between approximately 50 and 75 s, depending on the thickness of the FRP repair system, to achieve acceptable results. Also, only Teflon sheets larger than 100 mm2 were detectable by the infrared camera. Areas less than 100 mm2, do not seem to significantly alter the heat flow and thus were not picked up by the infrared sensor (see Figure 2.5). On these areas, although the precise shape of the Teflon plate is hard to recognize and existence of flaws cannot definitely be confirmed, some small heat disturbances can be seen at these locations. 19 Figure 2.6 Left: Thermographic image of trial slab; Right: Teflon sheet arrangement underneath FRP. As it can be seen from Figure 2.5, locations at which Teflon plates are embedded to simulate areas of debonding emit more heat than areas with perfect or near perfect bond after a 65 s heating period. From this thermograph, areas of fiber overlapping can also be detected; these generate a grid-like pattern. The smaller glowing yellow spots are ”hot spots” generated by interruption of heat flow caused by the microscopic gaps between concrete and FRP that act like an isolator layer. In other words, heat does not flow into the concrete and remains in the FRP shortly after the cooling process has started. 20 3 IMPACT-ECHO Impact-echo is a non-destructive method of material testing that records the travel time of impact generated transient stress waves using a sensitive broadband transducer and thus provides valuable information about continuity and integrity of subsurface material. Developed in 1985 [22], the method uses steel ball bearings as a mechanical impact source to generate low-frequency short-duration stress waves with sufficient energy to penetrate concrete. In this method the transient stress waves generated are reflected by boundaries of discrepancy in sound wave speed, thus allowing for detection of material discontinuities such as cracks, voids, or delaminations. 3.1 Fundamental theory The impact-echo technique is developed based on the theory that the impact of a metallic object on the surface of a specimen leads to generation of stress waves in the material. Stress waves can generally be generated by any mechanism capable of producing a force that varies rapidly with time, such as the piezoelectric transducer of a portable ultrasonic non-destructive digital indicating tester (PUNDIT) or simply collision of two solid bodies. In order to have a well-defined and mathematically simple input, which in return generates waves with characteristics that facilitate signal interpretation; steel ball bearings are adopted as the preferred source of impact. These ball bearings, made of hardened steel and ranging from 3 to 19 mm in diameter, are attached to a spring mechanism. Echo readings from the impact are recorded by a piezoelectric transducer-receiver and then digitized. The range of impact frequencies received is highly dependent on the impact time and directly proportional to the diameter of the impactor. Smaller impactors result in a broader signal frequency range, allowing for detection of reflected sound wave cause by material discontinuities. 21 There are three main type of waves generated after an impact, dilateral waves (P-waves), shear waves (S-waves), and Rayleigh waves (R-waves). Although propagation of transient stress waves in a heterogeneous bounded solid, such as an element of a concrete structure, is a complex phenomenon, these movements can be simplified. This can be seen in Figure 3.1, which provides a cross-sectional schematic that explains the direction and mode of propagation of each wave in a plate-like concrete specimen at δt time after an impact. Figure 3.1 Schematic representation of P-, S-, and R-wave fronts generated by an elastic impact. P-waves and S-waves are somewhat similar as they both travel in spherical wave fronts within a solid. They differ in the direction of solid particle motion. In P-waves the particle motion at the wave front is parallel to the direction of propagation; in S-waves 22 particle motion is perpendicular to the direction of propagation, which produces shear stresses. Rayleigh waves that are surface waves, however, do not follow a spherical propagation pattern and are transmitted across the surface of the specimen in a circular manner. P-waves could be either compressive, in which particle motion is outward along the wave front, or tensile, in which particle motion is inward along the wave front. Once impact has occurred, the P-wave travels into the medium and is reflected by all surfaces opposite to the impact surface. Arrival of the reflected P-wave at the impact surface produces a small downward displacement, leading to completion of one cycle of P-wave after impact. As P-wave cycles continue, the amplitude of the P-wave is decreased with time because of the spherical spreading at the wave front and the corresponding wave energy dissipation and loss. R-waves, on the other hand, do not reflect off any edges as they travel across the impact surface and have no cycles. R-waves provide information about the duration of impact, which allows for determination of impact frequency. R-waves and P-waves are the only waves that are recorded by the impact-echo transducer. Through measuring the impact stress waves’ travel time, the impact-echo technique provides valuable information on the continuity of the medium in which wave propagation is accommodated. Flaw detection using impact-echo relies upon the detection of transient resonances caused by multiple reflections of P-waves between two interfaces. Figure 3.2 illustrates the impact-echo testing process. 23 Figure 3.2 Impact-echo testing setup. Note that as the impact is generated by the impactor, surface displacements are picked up by the transducer. 3.2 Signal analysis Figure 3.3 shows a cross-sectional side view of a concrete specimen equipped with an impact-echo transducer at a fixed location on the impact surface. It can be seen that the compressive P-wave is travelling in a spherical pattern shortly after impact and the R- wave is travelling across the surface. Once the impact-generated compressive P-wave passes the transducer, it produces a small upward displacement normal to the surface, resulting in generation of a positive voltage, as shown in the voltage versus time graph in Figure 3.3. 24 Figure 3.3 Schematic showing the movement of stress waves shortly after impact at time t, with corresponding ideal time vs. voltage graph. Notice the P-wave passing by the transducer. The R-wave always reaches the transducer after the P-wave, as the velocity of the R- wave is half the velocity of that of the P-wave. Passage of the R-wave across the surface induces a sharp downward displacement at the wave front, leading to generation of a negative voltage by the transducer, which is often part of the signal with the highest amplitude, as shown in Figure 3.4. 25 Figure 3.94 Schematic showing the movement of stress waves at time t + e, with the corresponding ideal time vs. voltage graph. Notice the R-wave passing by the transducer. Since the R-wave produces a sharp downward displacement shortly after the passage of the P-wave, the specimen surface recovers to zero position, resulting in no voltage generation and return of the voltage versus time graph to its original zero-voltage level. Reflection of the P-wave off the specimen boundaries is a tensile P-wave. Once this wave is received by the impact-echo transducer, the negative displacement caused by this wave produces a negative voltage signal, as seen in Figure 3.5. 26 Figure 3.5 Schematic showing the movement of stress waves at time t + 2e, with the corresponding ideal time vs. voltage graph. Notice the reflected P-wave passing by the transducer. The P-wave repeats this periodic motion with intervals between successive wave arrivals before complete dissipation. This time interval is described by Equation 3.1 and shown on the voltage–time diagram in Figure 3.6. 27 Figure 3.6 Time vs. voltage graph, showing the periodic movement of the P-wave in the specimen. P 2 C d t = (Equation 3.1) Where, t is the time interval, d is the thickness of the specimen, and CP is the speed of the P-wave inside the medium. 3.3 Impact-echo transducer The impact-echo transducer, developed by Thomas Proctor in 1982, is a small conical piezoelectric element attached to a brass backing [23]. It was designed to respond to only small displacements normal to the surface. It requires a thin sheet of lead foil between the transducer tip and the specimen surface for optimal coupling and unlike “pulse-echo” transducers does not require any coupling gel for contact. The Proctor transducer 28 generates positive voltages for upward displacements and negative voltages for downward displacements, allowing for acquisition of a voltage versus time graph. As the amplitude of P-wave induced displacements is at maximum and that of the S-wave is at minimum directly beneath the source of impact, placement of the transducer at a location adjacent to the impact point maximizes the effects of P-waves and minimizes those of S- waves. This is the primary reason for reliance of the impact-echo technique on propagation and reflection of P-waves rather than S-waves. 3.4 Wave speed Measuring the impact P-wave speed is vital in impact-echo testing. Wave speeds in homogeneous, semi-infinite, elastic solids are functions of Young’s modulus of elasticity, the mass density, and the Poisson’s ratio. The speeds of P- and S-wave propagating in finite solids are given by the following expression: )21)(1( )1( P υυρ υ −+ − = E C (Equation 3.2) )1(2S υρ + = E C (Equation 3.3) where E is Young’s modulus of elasticity, ρ is mass density, υ is Poisson’s ratio, CP is the P-wave speed, and CS is the S-wave speed. As the mechanical wave travel is made possible by micro-displacements of material particles, denser mediums accommodate higher wave speeds within. As seen in Equation 3.1, in the impact-echo technique, specimen dimensions and depths of material anomaly are linear, expressed as functions of P-wave speed CP. Therefore, the accuracy of test results is in part directly related to the accuracy of the P-wave speed measurement. P- wave speed can generally be found by one of two methods: (1) measuring the travel time 29 of P-waves between two transducers a know constant distance apart on a specimen surface and (2) carrying out the impact-echo test on a plate-like structure with known dimensions and then calculating the wave speed using Equation 3.1. The P-wave speed can also be derived by a third method: using measurements of R-wave speed and the Poisson’s ratio of the material of interest. However, because of the great probability of error associated with ascertaining the Poisson’s ratio of the material, this technique is rarely considered as an established method of P-wave speed measurement. 3.4.1 Direct method of P-wave speed measurement The P-wave speed can be directly measured using a dual transducer system, shown in Figure 3.7. In a dual transducer setup, the transducer units are rigidly held in a spacer bar. The spacer bar keeps the distance between the two transducers constant at all times. Figure 3.7 A dual transducer, necessary for measuring P-wave speed. 30 Although solid particle displacement at the P-wave front is parallel to the direction of wave movement and hence parallel to the surface of the specimen, owing to Poisson’s effect, near surface particle displacements are normal to the specimen surface. Therefore, these waves are detectable by Proctor’s displacement transducer. The suggested distance between the point of impact and the nearest transducer on the dual setup is 150 mm. Provision of this distance is necessary to allow enough time for the P-wave to separate from the S- and R-waves before reaching the first transducer. This distance also ensures that the P-wave generated displacements arriving at the second transducer have enough amplitude and can be easily identified. The accuracy of P-wave speed measurement is a function of digital data sampling intervals and the distance, L, between the two transducers. Theoretically, a longer L provides more accurate results, but because of attenuation in the wave front there is a practical constraint on the distance L as the particle displacements become too small to be detected. Studies have shown that the optimum distance between the two transducers should be approximately 300 mm [24]. Once the distance between the fixed transducers (L) is set, measuring the precise arrival times of the spherical P-wave fronts at the two transducers in microseconds allows, the P- wave speed CP to be calculated using the following equation: 12 P tt L C − = (Equation 3.4) where L is the fixed distance between the two transducers and t1 and t2 are precise arrival times of the spherical P-wave front at the two transducers. This procedure to measure P- wave speed is explicitly outlined in ASTM Standard C1383-98a [23]. 31 3.4.2 P-wave speed measurement from impact-echo testing on plates with known dimensions P-wave speed can be easily measured by performing an impact-echo test on a plate-like structure with a known thickness. Once the test is carried out, the peak signal frequency is used in the following expression to find the P-wave speed (CP). β/2P dfC d= (Equation 3.5) In the above expression, fd is the frequency at the maximum signal amplitude, d is the known thickness of the plate structure, and β is the shape factor, equal to 0.96 for plate- shaped structures. The accuracy of P-wave speed measurement in this method is influenced by the accuracy with which the plate’s thickness is known and by the digital sampling parameters used in a specific test setup. 3.5 Transformation from time to frequency domain Surface displacements caused by stress wave reflections off various internal and external flaws within a bounded solid often have a wide range of frequencies and amplitudes. Material flaws generate wave forms that are often too complex to be analyzed. Consequently, it is usually extremely difficult, if not impossible, to distinguish signals that are received from multiple flaws at different depths. To identify the precise wave arrival times and determine the key frequencies, the wave forms are transformed into the frequency domain using a Fourier transform. This process, which is done instantly after testing by the data acquisition computer, presents the important frequencies as peaks in the amplitude spectrum. The Fourier transform is based on the principle that any time- dependent function can be represented as a sum of sine curves of different amplitudes and frequencies. This transformation is carried out numerically in impact-echo testing on a digitized waveform using a method known as fast Fourier transform (FFT). FFT is the 32 backbone of the impact-echo data acquisition software, Impact-E, that transforms the voltage–time function to amplitude spectrum. Although only P-waves provide information of interest about a structure’s dimensions or presence of flaws, the displacement caused by the R-wave in a waveform can have a significant effect on the corresponding amplitude spectrum. R-wave displacements in an impact-echo waveform caused by the propagation of the R-wave, which are detected by the Proctor transducer, are seen to be a mirror image of the force–time function of the impact. Therefore, when an impact-echo waveform is transformed into the frequency domain, the resulting amplitude spectrum includes the frequency distribution of the force–time function of the impact. 3.6 Interfacial wave reflections and refractions Materials inspected by the impact-echo technique could be very similar, such as concrete and cementitious repairing material, or different, such as concrete and air. By definition, an interface is the boundary between two materials having different acoustic properties, namely, acoustic impedance (Z). Acoustic impedance is defined as wave speed times the density of material. Differences in acoustic impedances of two materials result in generation of the acoustic interfacial boundaries. ρ.PCZ = (Equation 3.6) Once a stress wave reaches an interface between dissimilar media, it undergoes both reflection and refraction. Although mode-conversion does not play a notable role in impact-echo testing, relative differences in acoustic impedance dictate the reflection and refraction of P-waves at an interface. Interfaces can be categorized into solid/air and solid/solid interfaces. 33 3.6.1 Solid/air interfaces The boundary between concrete and air is the most common interface in impact-echo testing. Although P-waves propagate within a medium in spherical wave fronts, as the stress wave undergoes multiple reflection cycles, the radius of the spherical wave front decreases and waves at a region around the impact point essentially behave as plane waves whose direction of motion is perpendicular to the impact surface. Therefore, movement of P-waves can be portrayed as ray paths and their reflective behavior analogous to that of light rays, as shown in Figure 3.8. Figure 3.8 Simplification of P-wave movement in plate-like structures, using the ray light analogy. 3.6.2 Solid/solid interfaces The boundary between concrete and fiber reinforced polymers is a perfect example of a solid/solid interface. Unlike the solid/air interface, where the majority of the incident stress wave energy is reflected, stress waves are partially reflected and partially refracted through the interface. In most analysis of the stress wave behavior, for simplicity it is assumed that (1) both materials in the interfacial zone are perfectly bonded and (2) the direction of incident wave travel is normal to the interfacial plane. Such assumptions 34 facilitate expressing the amplitude of the refracted and reflected waves as a function only of the difference in acoustic impedance of the two media separated by the interface. Equations 3.7 and 3.8 express the amplitudes of the reflected and refracted stress waves in terms of acoustic impedance. )( )2( 12 2 ifractedRe ZZ Z AA + = (Equation 3.7) )( )( 12 12 iflectedRe ZZ ZZ AA + − = (Equation 3.8) where Z1 is the acoustic impedance of the region in which the incident wave propagates in before reaching the interface, Z2 is the acoustic impedance of the region in which the refracted wave travels in beyond the interface, and Ai is the amplitude of the particle motion in the incident wave. If the quantity Z2 – Z1 in Equation 3.8 is negative, a phase change of stress wave is observed upon reflection. In other words, it can be said that a phase change occurs when the direction of particle motion at the wave front is reversed at the interface; hence, compression waves are reflected as tensile waves and vice versa. In impact-echo testing, the conditions at solid boundaries are often such that Z2 << Z1, which results in a stress wave phase change. Such phase change results in AReflected approaching –Ai and reflection of nearly the entire incident beam. When testing steel reinforced concrete, as the acoustic impedance of steel is much higher than concrete (Z2 >> Z1), the amplitude of the reflected stress wave is equal to that of the incident wave. It also must be noted that in such cases there is no phase change and the amplitude of the refracted wave is twice that of the incident wave. 35 When testing concrete with externally applied reinforcing/repairing FRP patches, Z2 is either equal or nearly equal to Z1. If Z1 = Z2, then there would be no wave reflection and all of the incident wave energy is transmitted. In this case, theoretically, reflections occur only when there is a material discontinuity such as the presence of cracks or unbonded interfacial areas. 3.6.3 Diffractions Diffraction waves are generated only when a P-wave is incident upon the sharp tip of a discontinuity or a crack. Diffraction waves propagate along cylindrical wave fronts centered on the sharp crack tip. Diffracted P-waves play a major part in determining the depth of surface opening flaws. 3.7 Spring force effects Stress waves are generated by mechanically tapping the steel ball bearings against the surface of the structure. The impact force generated varies from an equivalent force of dropping the ball bearing from a height of 0.2 to 4.0 m. In general, impact is characterized by the duration of contact time of the impact, the size of the ball bearing, and the kinetic energy of the ball bearing at impact. The relationship among these parameters is explained by the Hertz theory of elastic impact. During an impact, a portion of the kinetic energy in the ball bearing is transformed into elastic wave energy in the concrete. The maximum force is proportional to the kinetic energy of the moving ball at impact, and the particle displacements in the resulting stress waves are proportional to this force. The contact time is a linear function of the ball diameter and is theoretically dependent on the kinetic energy. The contact time of a steel ball dropped onto a normal strength concrete surface from a height h (metres) can be approximated by Equation 3.9; 36 1.0i 0043.0 h D t = (Equation 3.9) where ti is the duration of contact in seconds and D is the diameter of the ball bearing in meters. As mentioned earlier, the height, h, which is related to the kinetic energy of the ball bearing spring just prior to impact, can be approximated from 0.2 to 4 m, resulting in a h0.1 value ranging from 0.85 to 1.15, which shows that the contact time is not very dependent on the drop height or, in other words, the kinetic energy of the ball bearing at impact. 3.8 Choosing the proper impactor The most important data output in impact-echo technique is the frequency of the impact- echo signal peak. If we define the maximum frequency of useful wave energy to be fmaximum = (0.8ti) -1, using Equation 3.8 when h0.1 is assumed to be 1, simplifies the maximum useful frequency as D f 291 max = (Equation 3.10) where fmax is in hertz and D, the spherical impactor diameter, is in metres. In other words, a ball bearing impactor with a 5 mm diameter is capable of producing useful frequencies up to approximately (291/5mm) = 58 kHz. Therefore, for a wave speed of 4000 m/s, the corresponding maximum depth of detection is (using equation 3.5) d = 0.96 x 4000/(2x 58) = 33 mm. When the impact-echo technique is used to detect delaminations, the probable depth of the flaw must be considered when choosing the impactor size. The advantage of using small impactors is that they put energy into stress waves at higher frequencies, bringing more detail about the structure near the impact surface. The disadvantage of using small impactors is the complication of the resultant system response, as the maximum useful frequency increases, wave inhomogeneities in concrete, such as microcracks, air 37 entrainment voids, and refraction/reflections caused by rebars, have a larger effect on the output signal. From Equations 3.5 and 3.9 we can approximate the minimum flaw depth (d min) that can be detected by the impact-echo technique as a function of the impactor diameter as follows: Dd 7min = (Equation 3.11) The above relationship states that an impactor with diameter D can be used to detect flaws only at depths greater than approximately 7D. Therefore, an impactor with a 10 mm diameter is capable of detection flaws only at depths 70 mm or greater. In other, words, the 10 mm impactor does not generate frequencies that would detect any material anomalies in the region between the impact surface and the depth of 70 mm. It is, however, recommend by those who developed the method that in cases where there is significant uncertainty about the proper size of the impactor, the test should be started using the largest impactor; reducing the impactor size through a trial-and-error testing process will allow the proper impactor size to be chosen. 38 4 SITES OF INTEREST AND METHODOLOGY 4.1 Sites of interest 4.1.1 Safe Bridge (Vancouver Island, British Columbia) As mentioned in Chapter 1, the Safe Bridge was the first of the three bridges used to test the efficiency of non-destructive testing. In some ways the Safe Bridge served as a laboratory for the other two bridges. Located relatively close to Vancouver, it can easily be reached by a short 90 min drive and a 50 min ferry ride from downtown Vancouver. The practicality of infrared thermography and impact-echo methods was extensively evaluated on the Safe Bridge, improving test implementation skills prior to application of these non-destructive testing techniques on the St-Étienne-de-Bolton and Leslie Street bridges. 4.1.1.1 Description and background Constructed in 1955, the Safe Bridge is one of the few precast concrete channel beam (PCCB) bridges still remaining in use on Vancouver Island. The Safe Bridge is located close to Cowichan Lake on Vancouver Island in the community of Youbo near Duncan, BC. The bridge spans a small creek that flows into Cowichan Lake and has two lanes for motor traffic and a pedestrian sidewalk (Figure 4.1). The clearance under the bridge is about 1.3 m at the upstream and 2.4 m at the downstream end [26]. Figure 4.19 Safe Bridge, Vancouver Island, BC. 39 The bridge is not on a heavy traffic route but it often must accommodate heavy logging trucks. As the bridge was designed in 1955 for H15 loading and design life of 50 years, well below the current Canadian Highway Bridge Design Code for CL-625 trucks and 75 year design life, it seemed vital to ensure the soundness and integrity of the structure. The Safe Bridge is constructed of 10 8-m-long PCCBs with an approximate width of 90 cm connected by shear joints with a total span of 8.2 m and a clear span of 7.6 m abutment-to-abutment. The total width of the Safe Bridge is approximately 9 m, which includes two motor traffic lanes and a 1.5 m pedestrian sidewalk. 4.1.1.2 PCCB bridges’ design deficiencies During a series of inspections in 1990s, it was determined that many of the PCCB bridges have undergone serious deterioration and are in need of either significant repairs or complete replacement. As all PCCB bridges were designed in accordance with the AASHTO code of the 1950s, which required only 1 inch of concrete cover over primary flexural reinforcements, the cover thickness in these bridges is insufficient according to today’s standards [27]. Consequently, a significant amount of reinforcement corrosion is generally exhibited in PCCB bridges, resulting in severe spalling and deterioration of the cover concrete. The other concern with the Safe Bridge design as a PCCB bridge was its low shear capacity. As it was built prior to 1970, the bridge’s original shear design before FRP retrofitting did not comply with newer bridge design standards [26]. The latter problem was of great importance, as shear failures in reinforced concrete beams are sudden and often catastrophic. 4.1.1.3 Retrofitting With the common financial constraints on infrastructure repairs, the owner, the BC Ministry of Transportation (BC MoT), became interested in the FRP repair technique. 40 An FRP strengthening system seemed ideal for the Safe Bridge’s problems. Not only did it improve the beam performance in terms of shear and flexure, but also the externally attached FRP acted as a hydrophobic layer of protection for the steel reinforcement within the concrete. Application of FRP to such a relatively large surface area proved somewhat time consuming and expensive when the surface preparation process necessary prior to the actual FRP application was considered. Consequently, in September 2001, it was decided that a newly developed technique, called sprayed fiber reinforced polymer be utilized for the first time in the repair. The sprayed FRP was inspired from shotcreting, the process of projecting concrete or mortar at high speed pneumatically onto a surface. Sprayed FRP consists of the use of a spray gun to shoot polymer and glass fibers concurrently. Resin/catalyst mixture from the lower nozzle of the spray gun and chopped fibers from the top-mounted chopper unit are sprayed at high speed simultaneously on to the repaired surface [28]. The combined streams transform into a two-dimensional randomly distributed fiber layer encapsulated by a catalyzed resin matrix. This process allows the gun operator to build up the FRP reinforcing layer to any desired thickness. The fiber length could be altered from 8 to 60 mm, depending on the application [26]. After spraying, a ribbed aluminum roller is used to force out any entrapped air voids and to work the material into a consistent thickness. One of the advantages of the sprayed FRP technique is that since the fibers are randomly distributed on the applied surface, the sprayed layer is theoretically two-dimensionally isotropic with identical mechanical properties in all directions. This advantage comes as a trade-off with loss of strength compared to the more traditional unidirectional continuous FRP laminates in their fiber alignment direction. Other advantages of sprayed FRP over traditional wrap FRP are more ductility and higher fracture toughness [26]. The sprayed FRP requires much less surface preparation, which results in cost and time savings. 41 Upon inspection of the Safe Bridge where testing was being carried out on flexural elements, two exterior girders (#1 and #10) were chosen, as they were most exposed to weathering. Girder #8 was chosen, as it was the boundary between the motor road and the sidewalk. Girders #2 and #5 were chosen arbitrarily. The girders’ numbering is shown in Figure 4.2. Each PCCB bridge girder has two legs for more bending capacity. For reference purposes, the left and right legs when looking at the cross-section facing east were named A and B, respectively, as seen in Figure 4.3. Figure 4.2 Section and plan view of the Safe Bridge showing the girder numbering and geometric dimensions. Figure 4.3 Sectional view of a channel girder from the Safe Bridge showing naming of the legs. 42 4.1.1.4 Climatic conditions The province of British Columbia is well known for its relatively mild climate. Compared to other Canadian provinces, BC has moderate winters with occasional subzero temperatures and an insignificant amount of snowfall along its coastline. As the seasonal temperature variation is relatively small in this province, the mild climatic conditions are also observed during the summer. Although rain fall levels are high in the coastal areas owing to the divertive role of the Rocky Mountains, use of de-icing salts for road maintenance during winter is rare and uncommon along the western coast of Canada and all of Vancouver Island. Figures 4.4 and 4.5 provide average daily temperature, daily extremes, and snowfall from 1971 to 2000 for the Cowichan Lake area [27]. Annual Temperature Profile -10 0 10 20 30 Ja n Fe b M ar ch Ap ril M ay Ju ne Ju ly Au g Se p Oc t No v De c MonthT e m p e ra tu re [ C ] Safe Bridge Figure 4.4 Safe Bridge's annual temperature profile. 43 Snow Fall Profile 0 5 10 15 20 25 30 Ja n Fe b M ar ch Ap ril M ay Ju ne Ju ly Au g Se p Oc t No v De c Month S n o w F a ll [ c m ] Safe Bridge Figure 4.5 Safe Bridge's annual snowfall profile. As, it can be seen from Figure 4.4, temperature is rarely below 0 °C; hence, structures in the Cowichan Lake area do not undergo many freeze–thaw cycles. The most probable period for snow fall is from November to December, and it is limited to a maximum of 26.6 cm, as seen in Figure 4.5. 4.1.2 St-Étienne Bridge (St-Étienne-de-Bolton, Québec) 4.1.2.1 Description and background St-Étienne Bridge is located near St-Étienne-de-Bolton, approximately 50 km west of Sherbrooke, Québec, spanning Autoroute 10. The bridge was constructed in 1962 at the same time as the formation of Autoroute 10. It is a two-lane bridge with a 1.2 m wide pedestrian sidewalk. The bridge is supported by a total of 18, 760 mm diameter circular columns, which are arranged in three rows at the two sides and at the mid-span. In 1997, nine of the columns that were relatively closer to the highway were retrofitted, some with cement mortar and all with wrap FRP reinforcement for protection against further corrosion of the steel reinforcements. Since the retrofitting project was included in a research project through the University of Sherbrooke [28]; various FRP systems were employed for a durability investigation in identical environmental conditions. 44 In addition to the retrofitting, four of the FRP reinforced columns were also instrumented and monitored using fiber optic sensors for displacements due to temperature variation, corrosion, and strains due to loading. Data acquired from these sensors showed highly unstable strain conditions due to corrosion. In this experiment, four of the original nine retrofitted columns were numbered 1 to 4 (as seen in Figure 4.6) and inspected using infrared thermography and impact-echo. The non- destructive results were compared with the semi-destructive direct mechanical pull-off test. Columns 1 and 3 were reinforced by carbon fiber reinforced polymer (CFRP) produced by Mitsubishi and Tonen; and columns 2 and 4 were reinforced by glass fiber reinforced polymer (GFRP) manufactured by Tonen and Tyfo, respectively. Figure 4.6 Plan view of the St-Étienne Bridge showing the numbering of the columns (note that the figure is not to scale). 4.1.2.2 Climate conditions The Canadian province of Québec is historically known for extremely cold winters and large annual temperature variation. Since continuous usage of de-icing salts during the 45 cold season is common for road maintenance, severe corrosion in bridge reinforcements and transportation supporting infrastructure is inevitable (Figure 4.7). Figures 4.8 and 4.9 show average daily temperature, temperature variation, and amount of snowfall from 1971 to 2000 [27]. Figure 4.7. Regular usage of de-icing salts induce a significant amount of corrosion on the St-Étienne Bridge. Annual Temperature Profile -30 -20 -10 0 10 20 30 Ja n F e b M a rc h A p ri l M a y Ju n e Ju ly A u g S e p O ct N o v D e c Month T e m p e ra tu re ( C ) St-Ètienne Figure 4.8. St-Étienne Bridge's average annual temperature profile. 46 Snowfall Profile 0 10 20 30 40 50 60 70 Ja n Fe b M ar ch Ap ril M ay Ju ne Ju ly Au g Se p O ct No v De c Month S n o w fa ll [ c m ] St-Ètienne Figure 4.9. St-Étienne Bridge's average snowfall profile. As can be seen from Figure 4.8, temperature is most critical from March to April and from October to November. It is during these two periods that structures go through multiple cycles of freeze–thaw. Freezing of water, within the pore structure of concrete, leads to volumetric expansion from hydraulic pressures and damage. The average temperature is subzero from November to March with the depth of snowfall exceeding 30 cm, providing sufficient basis for the assumption that during these months, de-icing salt is regularly applied to prevent vehicle skidding and traffic accidents. 4.1.3 Leslie Street Bridge (Toronto, Ontario) 4.1.3.1 Description and background The Leslie Street Bridge, built in the 1960s, is located in Toronto, Ontario, and accommodates Highway 7. The highway has seven westbound and six eastbound lanes at this location. The columns are approximately 920 to 1010 mm in diameter. The specified concrete strength was 20.7 MPa (3000 psi). After decades of overuse and neglect, the 47 safety of the aging bridge was critically questioned. Due to extensive usage of chlorides on Highway 7 and Leslie Street during winter, severe corrosion was seen in the reinforcing steel embedded within the concrete structures and the steel girders. Figure 4.10 shows an example of severely corroded bridge columns. Figure 4. 10 Leslie Street Bridge. Note the level of corrosion in steel flexural members. The concrete is spalled off and the exposed steel reinforcement is corroded, neither of which are uncommon. As external application of FRP wraps provided promising results for the bridge’s on-going corrosion problems, all of the bridge’s columns were retrofitted using GFRP wraps in 1996. Figure 4.11 shows the columns before repair with GFRP in 1996. The condition of the repaired column in 2008 after exposure to extreme environmental conditions for over 12 years shows excellent performance of FRP. 48 Figure 4.11 Leslie Street before and after the application of FRP repairs. To better understand the integrity of the FRP bond on this bridge, two columns between the first and second spans were tested. The first column on the northern end was chosen as the exposed specimen and the fourth column from the northern end was chosen to represent one of the least exposed protected specimens. Figure 4.12 shows part of the retrofitted bridge where the direct pull-off tests were carried out. Figure 4.12 Leslie Street Bridge. Columns on which the direct pull-off test was performed are marked. 49 4.1.3.2 Climatic conditions The Canadian province of Ontario has a harsh climate. Similar to Québec, it has extremely cold winters and equally hot summers. Continuous usage of de-icing salts during winter is common for road maintenance and severe corrosion in bridge reinforcements and transportation supporting infrastructure is commonly encountered throughout the province. Figure 4.10, shows the effects of the long-term use of de-icing salts on the Leslie Street Bridge. Figures 4.13 and 4.14 show average daily temperature and amount of snowfall from 1971 to 2000 [27]. Annual Temperature Profile -10 -5 0 5 10 15 20 25 30 Ja n F e b M a rc h A p ri l M a y Ju n e Ju ly A u g S e p t O ct N o v D e c Month T e m p e ra tu re ( C ) Leslie St. Bridge Figure 4.13 Toronto's annual average temperature profile. 50 Snowfall Profile 0 5 10 15 20 25 30 35 40 45 Jan Feb March April May June July Aug Sept Oct Nov Dec Month S n o w fa ll [ m m ] Leslie Street Figure 4.14 Average annual snowfall profile in Toronto. 4.2 Methodology and experimental program The purpose of the test program was to determine the suitability of potential non- destructive techniques, infrared thermography and impact-echo, in assessing the long- term bond durability. The integrity of the FRP–concrete bond is difficult to determine, and there are numerous non-destructive testing techniques that claim to be capable of such evaluation. In order to evaluate the practicality and limitation of these non- destructive methods in the field, a series of experimental bond evaluations was carried out. The non-destructive test results were then compared with a quantified value acquired by performing a semi-destructive and relatively accurate mechanical pull-off test as a complementary testing technique. The bond durability was defined by the tensile strength between the FRP reinforcement layer and concrete substrate. 4.2.1 Experimental program The mechanical part of the test was carried out in accordance with ASTM C1583-04 [31] standard procedure (Figure 4.15). FRP reinforced structural elements under consideration 51 were first stripped of their protective paint and plastic barriers, wherever such barriers existed, with extreme care so as not to damage or induce an impact on the material. The process was carried out manually using a mallet and a chisel. These protective layers are generally placed on top of FRPs to protect the polymer from UV rays and direct exposure to precipitation and de-icing salts. Figure 4.20. Schematic of testing apparatus. After gridding the FRP surface, the locations of cores were arbitrarily chosen and marked as seen in Figure 4.16. For columns, the core markings were made on two grids, a grid exposed to the road with a 30° angle from the axis of the road and a grid on the opposite side of the column, as seen in Figure 4.17. The grids were drawn 635 mm above the ground, measured from the bottom of the grid. Cores taken from flexural members were taken from girder sides near the supports where shear is at its maximum and in some cases also at the mid-span. 52 Figure 4.21. Gridding and random selection of core locations. Figure 4.22. The position of the column grids with respect to the road. 53 Following this procedure, these locations were heated using a dual 500 W halogen lamp for 60 to 80 s (see Figure 4.18) and thermographic images were taken. The heating process is necessary for acquisition of thermal images, as detection of flaws and de- bonded areas has to be done by active thermography. Figure 4.23. Heating of FRP surface as part of active thermography. After carrying out the non-destructive portion of the experiment and determining the location of the cores, coring commenced as part of the direct pull-off test. Cores had to be deep enough to penetrate a few millimetres into the concrete and up to a maximum depth of 10 mm. For optimum results, coring was carried out in 30 s intervals with a 2 min gap between a pair of intervals to avoid localized over-heating of the FRP. The coring drill had an inner diameter slightly larger than the steel disks for a total area of 2284 mm2. The coring setup is shown in Figure 4.21. Rigid steel disks with 50 mm diameter rigid plates and concentrically threaded pullout extensions were then attached by a chemical adhesive. These adhesive materials are generally two-component epoxies that yield a bonding strength of 4 to 5 MPa. To ensure proper adhesive bond, the disks were mechanically secured for a minimum of 24 h (see Figure 4.19). After the epoxy was set, the disks were pulled-out using the test setup seen in Figure 4.20, which consists of a mechanically driven arm connected to a calibrated 54 dynamometer. The load was applied at a rate of 0.005 MPa/s and the maximum pull-off load was recorded. Finally, the core areas were filled with compatible materials and painted over. In some instances (especially in the case of the Leslie Street Bridge) a premature adhesive failure occurred in the epoxy forming the interface between the FRP and the steel disk. In such cases, the load recorded was used to calculate the strength of the FRP– concrete bond. The logic was that the FRP–concrete interface in question had at least supported the recorded load. Figure 4.24. Securing test disks during epoxy setting. 55 Figure 4.25. Pull-off testing using a Swiss-made DVNA Haftprufer Z16. Figure 4.26 The coring setup. 56 5 TEST RESULTS AND DISCUSSION 5.1 Non-destructive tests Two methods of non-destructive testing techniques, infrared thermography and impact- echo, were carried out on the Safe and St-Étienne bridges prior to a direct mechanical pull-off test. The non-destructive evaluation of the Leslie Street Bridge was limited to only impact-echo testing. The purpose of the direct mechanical pull-off test, as a semi- destructive test, was to corroborate the results attained from infrared thermography and impact-echo tests and to provide quantitative data for FRP–concrete bond strength. 5.1.1 Infrared thermography An active mode EATF was used to inspect the FRP–Concrete bond at critical locations near the supports and in the middle span on five of the ten bridge girders of the Safe Bridge and on both exposed and protected sides of the columns of the St-Étienne and Leslie Street bridges (see Section 4.2.1). For thermography, a FLIR-EX300 camera was used with a 25° interchangeable field of view, a thermal sensitivity of less than 0.8 °C at 25 °C, and a spectrum range of 7.5 to 13 µM. The thermographic images taken from the three structures are extensively presented in Appendix B. 5.1.1.1 Safe Bridge, British Columbia Infrared thermography was used to guide the approximate location of the cores for the direct pull-off test on this bridge. The average nominal thickness of the sprayed FRP layer was measured to be 7.3 mm from all three locations on the bridge; hence, the sprayed FRP surface was heated for approximately 70 to 85 s using a dual 500 W halogen lamp. This heating length, slightly longer than the optimal length found in the in-laboratory experimentation (see Section 2.3.2), was deemed necessary to diffuse sufficient heat to penetrate the entire sprayed FRP layer. As a result, the surface, temperature was relatively higher, which at times 57 required test repetition for better results. It was found, however, that repetition of thermography was required only in areas where the sprayed FRP layer thickness was more than approximately 10 mm. The thermography of the bridge showed bond deterioration in nearly all of near support areas, except for girder #2, due to intrusion of water between the FRP and concrete layers. This water intrusion can be explained by the unique geometry of the PCCB girders and their joints with the road; this is further explained in Section 5.3.2. The level of water intrusion varied from girder to girder, but the exterior girders #1 and #10 as well as girder #8 were revealed to be most affected by water intrusion. Excessive intrusion of water in girder #1 seemed to have enabled growth of an unknown type of anaerobatic fungi, which was clearly visible in the thermographic image of the FRP–concrete interface (see Figure 5.1a). A circular and a rectangular strip of SFRP delamination was detected at the top and bottom sides of girder #10, as seen in Figure 5.1b. Middle span inspection of girder #5 showed a circular flaw to the left of the fiber optics on leg B and a large rectangular area of sprayed FRP irregularity and inflation. 58 Figure 5.1 (a) Growth of fungi detected by thermography. (b) A circular and rectangular demination found on the side of the girder. (c) Detection of water pathway and fiber optics by thermography. Note the location of a pull-off test which indicated zero bond strength. (d) The water seepage pathway clearly visible in the thermographic image. The bond integrity seemed more intact in the middle span. Water intrusion underneath FRP at the middle span was minimal in all girders except for girder #10, where the water intrusion pathway was clearly visible in the thermographic images. The pathway of water seepage is noticeably distinguishable in the middle-span thermographic image of girder #10 presented in Figures 5.1c and d. Fortunately, routine water intrusion pathways were easily detectable because of rainy conditions on the day of testing. Seepage of water into the sprayed FRP–concrete interface was further investigated by the scanning electron microscope (SEM) to corroborate the thermographic results (see Section 5.3.2). 59 Owing to severe level of water intrusion, it was not clear whether debonding was initiation solely by intrusion of water or if it was triggered by mechanical straining followed by water infiltration. It should be noted that thermography clearly showed the inferiority of bond and areas where water has infiltrated at most locations on the Safe Bridge. 5.1.1.2 St-Étienne-de-Bolton Bridge, Québec Thermography on the St-Étienne Bridge yielded more valuable experimental results than the Safe Bridge in correlating bond strength and thermography, as no water seepage was detected in the interfacial plane and hot spots were simpler to analyze. The thermographic images clearly showed areas of FRP delamination on all columns as well as areas with more superior bond quality. The heating period was roughly 70 s and infusion of the heat in the material was relatively uniform. Figure 5.2 shows a thermographic image taken from column 1 in which values of measured bond strengths from the pull-off tests are presented. Notice that thermography indications of debonding corresponded well with lowering of the bond strength. Unfortunately, this correlation is not precise, as thermography failed to show both small as well as deeply embedded debonding. It is presumed that the hot spots detected by the infrared camera are somewhat larger than the actual debonded areas. As the diffusion of input heat is interrupted by the presence of flaws in active thermography, some of the entrapped heat is transferred to the flaw boundaries, generating a thin heat layout in the themographic image of the FRP surface around the debonded area. This heat layout cannot be differentiated from actual delamination by an infrared camera. Cores near this layout zone indicate significant bond strength losses. On the other hand, cores that were entirely over-lapped with hot spots did not show any bond strength as the core fallout was inevitable. Partial strength loss in cores from the layout zone is due to stress concentration in bonded areas when some percentage of interfacial area of the core coincides with debonded FRP and bond strength elimination is a consequence of complete interfacial delamination. Figure 5.3 schematically shows doubling of stress on the bond due to the presence of 60 debonded areas when it is assumed that the bond strength is equal at all locations. In other words, if the bond strength is greater than zero, then the load required to pull-off steel disk A is 50% of the load required to pull-off steel disk B. Figure 5.2 Thermographic image of column 1 (shown on right) with bond strength values acquired by the direct pull-off test. Figure 5.3 Doubling of stress on the bond due to presence of debonded areas. 5.1.2 Impact-echo Impact-echo testing was carried out on all three bridges. As mentioned in Chapter 3, the impact-echo technique provides valuable information about integrity and internal 61 condition of subsurface material. Recently, usage of this technique for bridge inspection has become very popular; however, little research has been carried out to investigate the technique’s capability to evaluate the FRP–concrete bond. For impact-echo testing a system of dual-transducer (See Figure 3.7), pistol transducer (See Figure 3.2), and accompanying data acquisition system from Impact-Echo Instruments, LLC was used. Impact-echo offers relatively accurate quantitative results, and its deep material penetration capability makes it much superior to infrared thermography as a non- destructive test method. However, because of its highly localized area of detection, the method is incapable of any comprehensive large-scale assessment. Therefore, evaluation of any structure by using solely the impact-echo technique is extremely time consuming, costly, and inefficient. The thickness of externally applied FRP could generally vary from 2 to more than 15 mm. From Equation 3.10 we know that even when a 2 mm diameter impactor sphere is used, the minimum depth detectable by impact-echo is 7 × 2 = 14 mm. Therefore, theoretically detection of any flaws at the FRP–concrete interface is highly unlikely when the FRP thickness is less than 10 mm. When the validity of this theory was put to the test, it was determined that when FRP delamination or deformation had occurred in the bridges, often no signal or occasionally ambiguous signals were received by the data acquisition unit; this was determined as unacceptable by the impact-echo program. Examples of an ambiguous signal and a signal with a clear frequency are presented in Figures 5.4 and 5.5 for better illustration of the concept. 62 Figure 5.4 An ambiguous impact-echo signal spectrum from location J0 on column 2 of the Leslie Street Bridge. The graph does not provide any significant information on bond condition nor in situ material discontinuity. The signal contains multiple peaks, corresponding to scattering of the impact displacement at the location of debonding. Locations with no delaminations and an intact bond showed more optimistic results, as the signal was acceptable, with the peak frequency showing the presence of internal cracks in concrete and the location of rebars. In these cases no signal disruption was seen from the interfacial boundary and this plane was not detected. An example of an impact- echo signal showing internal concrete flaws and the location of rebars is shown in Figure 5.5. 63 Figure 5.5 An impact-echo signal spectrum from location J6 on column 2 of the Leslie Street Bridge. The graph clearly shows the presence of in situ steel reinfocement at 43.5 cm below the impact surface. The spectrum signal shows a peak frequency of 4.9 kHz. Using Equation 3.5 with β = 0.96 and a measured impact wave speed of 4444 m/s, this frequency corresponds to a wave reflection at approximately 4444*0.96/(2 x 4.9) = 435 mm below the impact surface, most likely caused by the presence of steel reinforcement bars at this location. As the nominal thickness of FRP applied on the Leslie Street Bridge was approximately 5 mm, the signal frequency that corresponds to the concrete-FRP interface at this depth should be 42.7 kHz. Figure 5.5 does not show any signal excitation at or near this frequency, therefore the FRP bond was intact with no material discontinuity at location J6 on Column 2 of Leslie Street. Thus, it can be concluded that without material discontinuity the interfacial zone of FRP and concrete, which is the jointing plane of two materials with distinct properties, is not detectable in the impact wave spectrum. The frequency spectrum output from impact-echo testing on the three structures are presented in Appendix C. 5.2 Semi-destructive mechanical pull-off test Once non-destructive testing of the bridge was completed and the structure was assessed based on the thermography and impact-echo results, the direct mechanical pull-out test 64 was carried out to acquire quantitative, accurate, and relatively reliable results. In the case of the Safe Bridge, the pull-off test was carried out in accordance with ASTM C1583-04 [31] standard procedure, as mentioned previously. The rationale behind correlating the pull-off load and the bond strength was that stress value at failure is equal to or greater than the actual bond strength. Therefore, it can be assumed that in cases of adhesive epoxy failure prior to debonding, the FRP–concrete bond strength is higher than the load applied to cause the epoxy to fail. The pull-off load readings from various locations were used to calculate the tensile bond strengths and are reported in Table 5.1. A pull-off load transfer area equal to the core disk area over which the epoxy gel was applied was used in the calculations. Notice that in Table 5.1 data are provided for both carbon and glass fiber reinforced polymer repairs. Recall that CFRP systems were implemented on the Leslie Street Bridge, GFRP on the Safe Bridge, and a mixture of both CFRP and GFRP on the St-Ètienne Bridge. Table 5.2 Tensile bond strenth from all three bridges. Mean Standard Maximum Minimum (MPa) deviation (MPa) (MPa) (Years) Safe Bridge Middle Span 1.45 0.76 2.56 0.48 Glass 8 Safe Bridge End Spans 1.06 0.88 2.61 0.18 Glass 8 Leslie Street Bridge 1.89 0.84 3.30 1.00 Carbon 11 St-Étienne Bridge (Column 1) 0.72 1.11 2.85 0.09 Glass 12 St-Étienne Bridge (Column 2) 2.53 1.48 4.74 0.88 Carbon 12 St-Étienne Bridge (Column 3) 1.94 1.63 4.54 0.12 Carbon 12 St-Étienne Bridge (Column 4) 3.60 1.34 4.95 0.57 Glass 12 Ultimate Pull-off Strength Age of repairType of fiber A high standard deviation in data indicates that there is large variability in the measured bond strengths at the four sites. The mean bond strength varied from as low as 0.72 MPa to as high as 3.60 MPa. 65 As seen from Table 5.1, when the 12 year old St-Étienne Bridge is considered, GFRP systems appear to demonstrate higher bond strengths than the CFRP systems. This is, however, contradicted by data from the Safe Bridge. This suggests that rather than the type of FRP, it is the workmanship, methodology of application, properties of the cement substrate, exposure conditions, and loading conditions that govern the bond. Overall, the FRP–concrete bond is sound at most places, but there are some locations where the bond has been compromised. 5.2.1 Pull-off results of flexural members (Safe Bridge) The locations of the cores were chosen based on the thermography and impact-echo results. To test the suitability and precision of thermography as a non-destructive method of bond assessment, some of the cores were taken from locations that were presumed to have an inferior bond and some were taken from areas that reflected the presence of an intact bond. In general, from the girders under inspection, two cores were taken from the side of the east girder end, two cores from the side of the west girder end, and one core from the mid-span girder soffit. The location of each core on each girder varied from that of other girders, making the evaluation more random. After completion of the coring process, the steel disks were pulled using a custom-made pull-off apparatus provided by the contractor that recorded the force used to detach the core from structure. This force was later divided by the area of the steel disk to calculate the bond strength. The bond strengths recorded from the Safe Bridge are presented in Tables 5.2 to 5.5. 66 Table 5.2. Pull-off test results from column 1 of the St-Etienne Bridge Test Location Ultimate Test Load [KN] Ultimate Bond Strength [MPa] Test Bond Failure Type and Location A15 Column 1 0.49 0.25 60% debonding; 40% delamination A16 Column 1 7.46 3.80 Epoxy B0 Column 1 5.60 2.85 10% air void; 15% Concrete; 75% debonding B2 Column 1 9.31 4.74 10% debonding; 90% concrete C11 Column 1 3.40 1.73 Concrete E12 Column 1 5.67 2.89 10% debonding; 90% concrete E1 Column 1 5.48 2.79 60% concerete; 40% FRP E3 Column 1 1.75 0.89 Mortar F16 Column 1 2.95 1.50 Mortar F3 Column 1 0.51 0.26 30% debonding; 70% delamination G0 Column 1 8.54 4.35 Epoxy G13 Column 1 3.87 1.97 Concrete G5 Column 1 1.73 0.88 Mortar Average 4.37 2.22 Table 5.3. Pull-off test results from column 2 of the St-Etienne Bridge Test Location Ultimate Test Load [KN} Ultimate Bond Strength [MPa] Test Bond Failure Type and Location A15 Column 2 0.31 0.16 Epoxy A16 Column 2 7.58 3.86 Concrete with 2 mm x 6 mm air void A4 Column 2 0.33 0.17 FRP; 30% FRP remains on the column B0 Column 2 1.51 0.77 Epoxy B10 Column 2 0.29 0.15 FRP D13 Column 2 0.00 0.00 Sample failure during coring D16 Column 2 0.00 0.00 Sample failure during coring E1 Column 2 4.01 2.04 Epoxy F12 Column 2 0.47 0.24 Debonding F3 Column 2 0.90 0.46 Epoxy F7 Column 2 0.31 0.16 Premature epoxy G16 Column 2 1.77 0.90 FRP H6 Column 2 0.37 0.19 Delamination Average 1.62 0.83 67 Table 5.4. Pull-off test results from column 3 of the St-Etienne Bridge Test Location Ultimate Test Load [KN] Ultimate Bond Strength [Mpa] Test Bond Failure Type and Location B0 Column 3 0.47 0.24 Mortar B17 Column 3 4.59 2.34 30% FRP; 70% concrete B2 Column 3 0.24 0.12 Mortar B4 Column 3 0.00 0 Failure during coring C11 Column 3 3.59 1.83 Epoxy C15 Column 3 6.68 3.4 100% concrete D3 Column 3 0.00 0 Failure during coring F11 Column 3 8.91 4.54 55% FRP; 45% concrete; 2 tiny voids F4 Column 3 1.28 0.65 50% FRP; 50% concrete G14 Column 3 6.07 3.09 40% FRP; 40% concrete; 20% void H10 Column 3 7.17 3.65 15% FRP; 15% air void; 70% concrete H6 Column 3 0.75 0.38 Mortar Average Average 3.97 2.02 Table 5.5. Pull-off test results from column 4 of the St-Etienne Bridge Test Location Ultimate Test Load [KN] Ultimate Bond Strength [MPa] Test Bond Failure Type and Location A3 Column 4 6.36 3.24 Mortar B17 Column 4 7.89 4.02 Epoxy B5 Column 4 6.13 3.12 Mortar C10 Column 4 0.00 0.00 Concrete D1 Column 4 1.12 0.57 Mortar D16 Column 4 9.72 4.95 Concrete G1 Column 4 3.30 1.68 Epoxy G14 Column 4 9.50 4.84 95% concrete; 5% air G16 Column 4 9.27 4.72 Epoxy G6 Column 4 6.70 3.41 Mortar H12 Column 4 7.30 3.72 Concrete H5 Column 4 9.27 4.72 Mortar Average 6.96 3.54 The middle span bond strength seemed more promising than the near support bond strength, as shown in the Table 5.6. The nominal FRP thickness of the mid-span soffit 68 varied from 3 to 10 mm with an average nominal thickness of 4 mm. FRP thicknesses measured from near support cores are greater than those from the middle span, with average nominal thickness being 9 mm. FRP thickness as much as 14 mm was also noted on girder #8. Table 5.6. Pull-off test results from the Safe Bridge mid-span Test Location FRP Thicknes s [mm] Ultimate Test Load [kN] Tensile Bond Strength [MPa] 1A Beam soffit 3 4.30 2.19 1B Beam soffit 3 5.02 2.56 2A Beam soffit 3 2.46 1.26 2B Beam soffit 4 1.76 0.90 5A Beam soffit 8 ~0 0.00 5B Near beam soffit 10 ~0 0.00 7A Beam soffit 3 0.95 0.48 7A Beam soffit 3 1.98 1.01 10A Beam soffit 3 1.99 1.02 10B Beam soffit 3 4.30 2.19 Average 4 2.85 1.45 Adhesive, between FRP and substrate Adhesive, adhesion fixture to FRP Tensile Bond Failure Type and Location Adhesive, adhesion fixture to FRP Adhesive, adhesion fixture to FRP Adhesive, adhesion fixture to FRP adhesive, adhesion fixture to FRP Adhesive, adhesion fixture to FRP Mainly cohesive, FRP Adhesive, between FRP and substrate Adhesive, between FRP and substrate Bond strengths as high as 2.6 MPa were observed in the middle span and areas near the eastern girder’s end, but this value did not exceed 1.5 MPa near the western girder’s end. The average bond strengths, excluding areas of zero bond strength, were 1.5, 1.2, and 0.9 MPa at middle, east end, and west end spans, respectively (see Tables 5.6 to 5.8).. 69 Table 5.7. Pull-off test results from the Safe Bridge eastern near support end span Test Location FRP Thickness [mm] Ultimate Test Load [kN] Tensile Bond Strength [MPa] 1A East side 6 0.95 0.48 1B East side 13 1.70 0.87 2A East side 5 1.35 0.69 2B East side 10 5.12 2.61 5A East side n/a n/a n/a 5B East side 10 ~0 0.00 8A East side 9 ~0 0.00 8B East side 10 ~0 0.00 10A East side 8 ~0 0.00 10B East side 7 ~0 0.00 Average 9 2.28 1.16 Mainly cohesive, substrate concrete Mainly cohesive, substrate concrete Not accessible for testing Adhesive, during test preparation Adhesive, between FRP and substrate Adhesive, between FRP and substrate Tensile Bond Failure Type and Location Adhesive, adhesion fixture to FRP Adhesive, between FRP and substrate Adhesive, between FRP and substrate Mainly cohesive, FRP Table 5.8. Pull-off test results from the Safe Bridge western near support end span Test Location FRP Thickness [mm] Ultimate Test Load [kN] Tensile Bond Strength [MPa] 1A West side 9 ~0 0.00 1B West side 11 ~0 0.00 2A West side 10 ~0 0.00 2B West side 7 3.00 1.53 5A West side 5 ~0 0.00 5B West side 10 ~0 0.00 8A West side 7 ~0 0.00 8B West side 14 ~0 0.00 10A West side 6 0.36 0.18 10B West side n/a n/a n/a Average 9 1.68 0.86 Tensile Bond Failure Type and Location Adhesive, between FRP and substrate Mainly adhesive, between FRP and substrate Not accessible for testing Adhesive, between FRP and substrate Adhesive, between FRP and substrate Adhesive, between FRP and substrate Adhesive, between FRP and substrate Adhesive, between FRP and substrate Adhesive, between FRP and substrate Adhesive, between FRP and substrate Locations with zero bond strength represent cores that simply fell out of place once the coring drill completely penetrated the FRP layer. There were only two such cores in the middle span, where the FRP thickness was relatively smaller. These types of test deficiencies were experienced more in the near support locations, as predicted by the thermographic images. Such bond deterioration could be related to continuous water seepage in the underside of the FRP layer, lack of substrate surface roughness at the time FRP was applied to the bridge, and damage in the FRP by the heat generated during the coring process. Concerning water seepage, given the geometry of the channel girders, 70 some water had undoubtedly seeped to the underside of the FRP spray and undergone freezing and thawing cycles, causing bond deterioration. It is imperative, therefore, that after application the FRP edges are sealed properly to prevent any infiltration of water. 5.2.2 Pull-off results from vertical weight support members The locations of cores were randomly chosen on the columns of Leslie Street and St- Étienne-de-Bolton bridges. As mentioned earlier, two surfaces normal to a 30° plane from the axis of the road were chosen. The more exposed surface faced the road in front of the column, whereas the posterior less-exposed surface was protected from the road. Once the boundaries of test area were determined, an 8 × 8 grid of 60 mm × 60 mm squares was drawn, with its rows labeled alphabetically from A at the bottom to H at the top. The columns were labeled numerically, starting from 0 to 7 on the exposed side and 8 to 17 around the circular columns on the protected side. Indiscriminately selected core locations were then marked with an X to guide the coring contractor. As mentioned in Chapter 4, four columns of the St-Étienne Bridge were tested. The results of direct mechanical pull-off tests carried out on the columns of the St-Étienne Bridge are presented in Tables 5.2 to 5.5. The ultimate bond strength varied significantly on each column. Similar to some of the cores from the Safe Bridge, a few cores from columns 2 and 3 of the St-Étienne Bridge fell out during the coring or shortly after penetration of the concrete substrate. As no water infiltration was detected underneath the FRP layer and proper surface preparation had been carried out prior to application of FRP, bond deterioration could be associated with one or a combination of the following: (1) Disintegration of polymeric matrix by the heat of coring; (2) Physical damage incurred from the snow impact during the snow blowing process; (3) Environmental effects such as freeze–thawing of mist in the FRP– concrete interface, downgrading of the cohesive characteristics of FRP over time, etc. All of these debonding theories are briefly discussed in Section 5.3. 71 The average bond strengths were 0.83, 2.22, 2.02, and 3.54 MPa for columns 1, 2, 3, and 4, respectively. The low average bond strength of 0.83 MPa for column 1 does not prove the bond inferiority of Mitsubishi’s CFRP, as most of the tests resulted in premature failure of epoxy at relatively low loads. Failure of FRP during coring on the St-Étienne Bridge occurred only when testing columns reinforced with CFRP (columns 1 and 3). Thus, it is imperative that the bond deficiencies seen in the Safe Bridge are independent of the type of FRP used and can be related to factors such as the amount of surface preparation, workmanship, properties of the cement substrate, geometry of the element being reinforced, and the amount of insulation against the environment. The highest bond strength found in all cores was 4.95 MPa on column 4. Failure of this core occurred in the concrete substrate, seen in Figure 5.6a, with the FRP–concrete interface undamaged, which reflects much stronger bond strength. Figure 5.6b shows test B0 with 100% failure of the FRP layer and bond strength of 2.85 MPa. This mode of failure suggests that the weakest plane in the near surface zone is in the FRP layer. Failure of FRP was the rarest mode of failure experienced throughout testing of the St-Étienne Bridge; therefore, reduction in FRP strength cannot be concluded from these tests. Core H5, taken from column 4 and seen in Figure 5.6c, shows a 100% failure of substrate mortar. This mode of failure was seen often throughout testing. The bond strength seen in the cores with a 100% mortar failure varied from 0.12 to 4.72 MPa. This variability shows the huge inconsistency in the strength of mortar applied as repair prior to application of FRP and its significance in bond durability and strength. Figure 5.6d shows a complex failure plane that is partially in the substrate concrete and partially in the FRP reinforcement. The bond strength values suggest that the bond is 72 strongest at locations where 100% substrate concrete fails and weakest at locations where the substrate was repaired with mortar prior to application of FRP. Figure 5.6. Modes of failure in a direct mechanical pull-off test: (a) failure of concrete substrate, (b) failure of FRP, (c) failure of substrate mortar repair, (d) combined failure of substrate concrete and FRP. Overall, the bond strength of CFRP reflected higher bond strength values and seemed more reliable than GFRP on the St-Étienne Bridge after nearly a decade of service. The bond strength values in all products, regardless of FRP type and their manufacturer, seemed to be within acceptable limits, proving that external application of FRP is a durable and reliable method of repair and strengthening of columns. The pull-off test carried out on the Leslie Street Bridge was not as satisfactory as the test performed on the St-Étienne Bridge. Because of some shortcoming on the part of the designated contractor, the disks were left bonded on the structure in sub-zero 73 temperatures ranging from –18 to –28 °C for nearly for two weeks. It is assumed that this delay, coupled with the extreme temperatures at the time of testing, resulted in premature adhesive failure of nearly all steel pucks on column 2. The pull-off test results on the Leslie Street Bridge are presented in Tables 5.9 and 5.10. Table 5.9. Pull-off test results from column 1 of the Leslie Street Bridge Test Location Ultimate Test Load [KN] Ultimate Bond Strength [Mpa] Test Bond Failure Type and Location A0 Column1 2.0 1.0 Epoxy and FRP A5 Column1 3.9 2.0 Epoxy C1 Column1 2.0 1.0 Epoxy E1 Column1 5.9 3.0 Epoxy E4 Column1 2.0 1.0 Epoxy and FRP I1 Column1 0 0 Failure during coring I4 Column1 0 0 Failure during coring A8 Column1 6.5 3.3 Epoxy B6 Column1 4.7 2.4 Epoxy D9 Column1 3.9 2.0 Epoxy D11 Column1 5.9 3.0 Epoxy and FRP F8 Column1 3.9 2.0 Epoxy and FRP H8 Column1 2.0 1.0 Epoxy J11 Column1 2.7 1.4 Epoxy and FRP Average 4.2 2.2 74 Table 5.10. Pull-off test results from column 1 of the Leslie Street Bridge Test Location Ultimate Test Load [KN] Ultimate Bond Strength [MPa] Test Bond Failure Type and Location A0 Column 2 2.0 1.0 Epoxy B4 Column 2 0.0 0.0 Epoxy D2 Column 2 0.0 0.0 Epoxy D5 Column 2 0.0 0.0 Epoxy F1 Column 2 0.0 0.0 Epoxy F4 Column 2 0.0 0.0 Epoxy H0 Column 2 0.0 0.0 Epoxy I2 Column 2 0.0 0.0 Epoxy A6 Column 2 0.0 0.0 Epoxy A8 Column 2 0.0 0.0 Epoxy B10 Column 2 4.7 2.4 Epoxy D8 Column 2 0.0 0.0 Epoxy F10 Column 2 0.0 0.0 Epoxy G7 Column 2 0.0 0.0 Epoxy H9 Column 2 0.0 0.0 Epoxy I7 Column 2 0.0 0.0 Epoxy Average 3.3 1.7 The average FRP bond strength for the Leslie Street Bridge was 2.2 and 1.7 MPa for columns 1 and 2, respectively. As mentioned earlier, owing to poor performance of adhesive epoxy, the accuracy and correctness of the recorded data cannot be guaranteed. Therefore no conclusion can be drawn from the bond strength values acquired. It is imperative that the direct mechanical pull-off test should not be carried out in extremely cold temperatures, as the effectiveness of the adhesive material is drastically reduced. 5.3 Possible causes of FRP–concrete bond deterioration The loss of strength and deterioration of the FRP–concrete bond can be related to many factors. The most noteworthy of the factors detected on the three structures in this study are briefly discussed in this section. 75 5.3.1 Insufficient surface preparation Application of FRP reinforcement, in general, requires a certain level of bond surface preparation. Depending on the severity of deterioration, surface preparation is more commonly followed by the repairing process. The surface preparation process can include one or a combination of surface grinding, sand blasting, sanding, and jack hammering to roughen the bond surface and water-jet washing, air blowing, or water rinsing to remove from the surface any organic materials, chemicals, and dust. Previous research has shown that failure to conduct an extensive surface preparation results in considerable reduction in bond strength, especially if sprayed FRP is being incorporated [29]. The record reflects minimal surface preparation techniques to the structure prior to application of sprayed GFRP. Hence, significant reduction of sprayed GFRP bond strength after eight years of service may be partially a result of insufficient surface preparation. 5.3.2 Water seepage Figure 5.7 is a schematic illustrating the presumed water seepage pathway within the concrete. It is very likely that water is accumulated between the asphalt fill and the curved girder corner. Owing to increases in size and weight of trucks and trailers since 1955 and the corresponding changes in bridge design codes, it can be concluded that the Safe Bridge is under-designed by today’s standards. Though no failure has occurred, stresses exerted by the modern motor traffic on the 54 year old concrete structures generated and widened cracks under load, allowing water to seep into the concrete. As GFRP is a impermeable material, it does not allow any water seepage, resulting in dripping along the GFRP and drainage through the areas not covered by GFRP. 76 Figure 5.9. The presumed pathway of water seepage under the sprayed FRP applied on the Safe Bridge. To further investigate water seepage and its presence behind the GFRP layer, a scanning electron microscope was used to randomly inspect some of the GFRP cores taken for the direct pull-off test for remnants of any salt or rust. EDX and X-ray scans (Figure 5.8) showed the presence of rust particles and salt crystals on the FRP underside on the posterior surface of the cores. 77 Figure 5.10. EDX and X-ray scans of the FRP underside from core 5A-east, showing presence of rust particles and salt crystals. As seen in Figure 5.8, rust particles are vividly distinguishable by their shining electron backscatter dispersed randomly all over the surface. Smaller salt particles, although not clearly visible in the SEM image, were found near the rust particles at all times. X-ray mapping of areas in close proximity to rust particles showed a uniform presence of salt. 78 Presence of salt near rust particles is proof of water seepage underneath the FRP reinforcement. Theoretically, water carries diluted de-icing salts from the road surface into the concrete structure and beneath the GFRP layer. As corrosion of steel reinforcement is imminent, corroded steel/rust particles are also carried by water behind the GFRP layer. After evaporation/drainage of water, salt and rust particles remain in the water seepage pathway in the GFRP–concrete interface. Hence, salt is consistently seen near rust particles in the interfacial surface. Water seepage underneath the GFRP reinforcement results in extensive bond deterioration by water flow at the interfacial plane and volumetric straining due to freezing of water during colder temperatures. According to Transports Québec, application of deicing agents depends on the temperature and desired level of melting. Sodium chloride is generally used for clearing residual snow left after maintenance operations in temperatures above –10 °C, with a small addition of calcium chloride in cases of temperatures below –15 °C. As these chemicals are known for infiltrating concrete, use of them initiates and accelerates the corrosion of steel reinforcements within the concrete, which is generally placed close to the surface for optimal mechanical performance. 5.3.3 Heat of coring Drilling cores in the FRP layer of the Safe Bridge proved to be a difficult undertaking. Because of the projectile nature of applying FRP, the thickness of the GFRP layer varied significantly across the structure. Though this layer did not exceed 12 mm at the middle span, thicknesses as much as 36 mm was seen near the supports. Penetration of such a thick GFRP layer using a drill produced a significant amount of frictional heat. Although the drilling process was carried out in intervals to reduce this effect it is possible that the heat of drilling deteriorated the FRP integrity and consequently its bond with concrete in areas where the FRP layer was substantial. The heat of drilling could be a fair justification to the sudden fallout of some of the FRP cores near the bridge supports as the underlying concrete was reached by the drill. 79 5.3.4 Snow clearing process Another factor that could possibly contribute to structural damage of the St-Étienne Bridge is impact attributed to the snow clearing process. As a snow-clearing vehicle passes underneath the bridge, the volume of snow accumulated in front of its blade is pushed to the sides, hitting the end-span columns. The impact force does not seem to be significant at the beginning of the cold season, but as the snow piles up on the sides of the road and around the bridge supports, the increase in the mass of the snow could lead to large forces. In a study carried out by in 1966 on the impact force of avalanches, a similar concept of snow moving with high velocities in a large mass is addressed [32]. This study uses Equation 5.1 as an approximation to calculate the average impact stress of snow. 3 2 3 22 3 2 3 200 400)( k ukukuk p ρρρ ++ = (Equation 5.1) where p is pressure, ρ is average snow density, u is snow velocity, and ρ45.53 =k . Assuming the impact force per unit volume of snow from a grader can be related to an impact load of an avalanche, with an average snow density of 0.15 g/cm3 [34], a maximum blower velocity of 60 km/h, and an air-to-ice ratio of 10% for compacted snow, the stress on the exposed bridge columns can be estimated to be 0.6 kN/mm2. Although this pressure is not large enough to cause instantaneous structural damage or failure, in time it due to its repetitious nature could have a significant cumulative effect on the strength of the bond between the applied FRP and the concrete surface. 80 6 CONCLUSIONS AND FUTURE RESEACH RECOMMENDATIONS 6.1 Conclusions Use of FRP to repair, strengthen, or retrofit is one of the finest solutions to the problem of deteriorating transportation infrastructure. Hence, a comprehensive knowledge of its durability and conditions of better preservation in the field is of high importance. Although there will always be semi-destructive methods of FRP bond evaluation, establishment of an efficient non-destructive testing technique as a reliable method of bond evaluation seems to be of great interest for economic reasons. In theory, many of the non-destructive testing methods seem suitable for bond assessment, but in practice they fail to provide feasible and practical results. In fact, they may be completely unsuitable. Durability assessments were carried out on three FRP strengthened bridges by conducting FRP–concrete bond tests. These structures represent a wide variation in severity of environmental exposure, length of time in service, type of strengthening, and type of FRP product used. They also encompass flexural and shear strengthening of beams and strengthening of columns. These bond tests included active infrared thermography, impact-echo testing, and the direct mechanical pull-off test. 6.1.1 Infrared thermography Thermography provides a good qualitative assessment of the bond, and the technique is extremely useful when inspecting vast areas. However, it cannot detect small changes in the bond quality, nor can it provide a quantitative assessment of the bond strength. 81 Thermography can only determine areas of complete delamination and debonding larger than 100 mm2. Thermography was performed on only two of the three bridges tested, the St-Étienne and Safe bridges. During inspection of the Safe Bridge, it was found that active thermography using halogen lights as the external source of heat is virtually ineffective when the FRP thickness is more than 10 mm. Therefore, depending on the type of infrared sensor and method of surface heating, the depth of detection can vary in active thermography. The depth of detection is relatively shallow and rarely exceeds 25 mm. Thermography is recommended for overall qualitative bond inspection provided that the implementation of thermography is complemented by direct mechanical pull-off tests at critical locations. Carrying out the pull-off test allows the inspector to generate a bond strength contour of the structure. 6.1.2 Impact-echo technique The impact-echo test was performed on all three structures. The method, in general, seemed very useful and fairly accurate in detecting internal flaws of subsurface concrete and the location of steel reinforcements. Unfortunately, the output signal peak can only be used in a limited way to acquire the exact location of internal flaws and cracks within a structure, and the method was incapable of properly evaluating near and subsurface zones as generation of suitable impact frequencies for this depth by the impact-echo setup was not physically feasible. As discussed earlier, the impact is only generated manually in this technique; hence, there are practical limitations on the size of impactor used. In order to evaluate areas near the concrete surface and the bond interface, frequencies between 40 and 1700 kHz are required. High-frequency impact sound waves are generated when only smaller impactors are utilized [33]. Experimentation showed that there is a size limitation on the smallness of the impactors, as extremely small impactor spheres are literally incapable of generating proper impact forces for testing purposes. 82 With a 2 mm diameter impactor, the test was successful when there was no debonding and a material discontinuity was present in the near surface zone and the interfacial plane. However, in areas where debonding was detected either by infrared thermography or the direct pull-off test, the method was unsuccessful and; either ambiguous and incomprehensible signals were recorded or no output signal was generated. Also, the impact-echo technique is not suitable for large-scale detection or for testing vast areas because the impact signal received corresponds only to a limited area on the surface. Therefore, although the implications of the impact-echo technique are very useful for in- depth inspections of structures, the test is not recommended for FRP bond evaluation at or near surface locations; i.e. in the case of extended strengthening using FRPs. 6.1.3 Mechanical pull-off test The mechanical pull-off test seemed relatively reliable and provided quantitative bond strength values. The method was used extensively as a complementary technique to assess the accuracy of non-destructive technique results. However, the technique has some downsides to it, namely the relatively long time required for carrying out the test, the damage induced on the FRP layer by the coring process, and the need for well trained and experienced operator. Also, during testing it became obvious that the adhesive epoxies currently available are very susceptible to moisture and temperature. Thus, the test has to be carried out only in optimum environmental conditions. The mean bond strength acquired by the direct mechanical pull-off test varied from as low as 0.72 MPa to as high as 3.60 MPa. The variability in bond measurements was inordinately high. The FRP–concrete bond seemed sound at most places in all three structures after, on average, a decade of service, but there are some locations where the bond has been severely compromised. 83 In the case of the sprayed FRP used on the Safe Bridge, it was observed that the bond is weaker in areas where the FRP is thicker. This phenomenon could be explained by two theories. (1) As the mass is more significant at these areas, the gravitational force on the material prior to hardening at early stages of application makes the FRP sag and harden in an undesirable form, detached from concrete. This is seen frequently at corners and curves. (2) As all the thicker areas were found near the bridge supports and the presence of water is much more severe in these areas, water seepage could have been a major player in bond deterioration and the thickness of the FRP is irrelevant. However, when considering all three structures, it is conclusive that rather than FRP type (glass or carbon) or the application method (spray or wrap), the workmanship, properties of concrete substrate, and structural detailing appear to be the important factors governing the bond. As seepage of water to the FRP underside severely reduces the bond strength, it is imperative that after application the FRP edges be sealed properly to prevent an infiltration of water. 6.2 Future research and recommendations Bond evaluation should preferably be carried out non-destructively, as repairing the tested areas in a semi-destructive test will be time consuming and expensive and in some instances comparable to complete replacement of FRP. As for the tests discussed in this document, the following were the factors which, if altered, would have enhanced the corresponding technique and its results. 6.2.1 Infrared thermography The active thermography on the three bridges was conducted using a dual 500 W halogen light as the external source of heating. The quantity of heat emitted from the dual halogen 84 light proved to be dependent on the atmospheric conditions and the position of the light with respect to the surface at the time of testing and, hence, inconsistent on daily basis. Although the amount of heat transmitted to the material could be calculated theoretically, these calculations do not seem to be reliable. The position and the angle of light projection change the diffusion distribution of the heat, making it inhomogeneous at times. To remedy this shortcoming, it is recommended that a more reliable source of heat that is also less susceptible to the environment be used. One promising example of such a heating source could be a heating blanket that can be wrapped around columns and underneath girders to heat the structure’s surface. These blankets use a series of embedded electric heating elements for heat generation, and as they cover the material surface, and provide better insulation to the surface. 6.2.2 Impact-echo test The impact-echo test provided valuable information about the interior conditions of the structure. However, operating through the use of sound waves, it is not recommended to use impact-echo for the purposes of near surface bond evaluations. Alternative techniques such as the use of microwaves in ground penetration radar (GRP) seem more promising for future research and might provide useful results. 6.2.3 Direct mechanical pull-off test The direct mechanical pull-off test was carried out successfully on the Safe Bridge and the St-Étienne Bridge; however, it did not provide highly reliable data on the Leslie Street Bridge. As mentioned earlier, this test deficiency at Leslie street bridge can be explained by the lack of technical experience of the testing team and extremely cold temperatures on the day of the testing. As the adhesive epoxies do not perform well in subzero temperatures, it is recommended that the direct pull-off test be carried out in milder temperatures, between 10 and 20 °C. A number of drilled cores failed prior to any pull-off test. Though the lack of the FRP– concrete bond was the main source of this phenomenon, increasing the diameter of the 85 steel disks used in a test from 50 mm to a larger dimension should provide better results by reducing the number of localized outliers and representing a larger area on the structure. This alteration could also reduce the probability of premature adhesive epoxy failure. A separate study on the effects that FRP applicators’ experience could have on long term bond quality also seems equally relevant; as most bond deficiencies are suspected to be an indirect result of an in-experienced applicator being chosen to apply the FRP on the structures. 86 REFERENCES 1. Nanni, A., “Relevant Field Applications of FRP Composites in Concrete Structures,” CCC 2001 – Composites in Construction International Conference, Porto, Portugal, October 2001. 2. Demers, M., Neale, K., and Sheikh, S., FRP Rehabilitation of Reinforced Concrete Structures, Manual No. 4, ISIS Canada, University of Manitoba, Winnipeg, Manitoba, 2008, 168 p. 3. Mufti, A., Banthia, N., Benmokrane, B., Boulfiza, M. and Newhook, J., “Durability of GFRP Composite Rods,” Concrete International, February 2007, pp. 37–42. 4. Balázs, G.L., and Borosnyói, A., “Long-Term Behavior of FRP,” Composites in Construction: A Reality: Proceedings of the International Workshop. Capri, Italy, E. Cosenza and A. Nanni, eds., American Society of Civil Engineers, 2001, pp. 84–91. 5. Banthia, N., Abdolrahimzadeh, A., Neale, K.W., Labossière, P., Demers, M., Mufti, A.A., Murison, E., Saltzberg, W., and Sheikh, S.A., “Bond Durability of FRP Repairs in Bridges,” Proceedings of the 4th International Conference on Structural Health Monitoring of Intelligent Infrastructure (SHMII-4), International Society for Structural Health Monitoring of Intelligent Infrastructure, ETH Zurich, July 2009. In press. 6. Akhtar, A., Nondestructive Evaluation History, MTRL486-Lecture Notes, The University of British Columbia, Vancouver, British Columbia, Fall 2009. 7. Malhotra, V.M., and Carino, N.J., Handbook on Nondestructive Testing of Concrete, 2nd Edition. ASTM International, West Conshohoken, 2004. 8. Feldman, R.F., Non-Destructive Testing of Concrete, CBD-187, National Research Council of Canada, Ottawa, Ontario, 1977, Available from http://irc.nrc- cnrc.gc.ca/pubs/cbd/cbd187_e.html [accessed November 8, 2008]. 9. Ohtsu, M., “Diagnostics of Cracks in Concrete Based on Acoustic Emission,” Nondestructive Testing, SP-112, H.S. Lew, ed. American Concrete Institute, Detroit, 1988, pp. 63–82. 87 10. Buyukozturk, O., and Rhim, H., “Radar Imaging of Concrete Specimens for Non Destructive Testing,” Construction and Building Materials, Vol. 11, No 3. 1997, Jones, R., Non Destructive Testing of Concrete, 1st Edition, 1962, Cambridge University Press, Cambridge. 11. McCann, D.M., and Forde, M.C., “Review of NDT Methods in the Assessment of Concrete and Masonry Structures,” NDT & E Internationl, Vol. 34, 2001. 12. Bouvier, C.G., “Investigating variables in thermographic composite inspections,” Materials Evaluation, Vol. 53, No. 5. May 1995, pp. 544–551. 13. ASTM Standard D4788-03, 2003, \"Standard Test Method for Detecting Delaminations in Bridge Decks Using Infrared Thermography\", ASTM International, West Conshohocken, PA, 2003. 14. Ghosh, K.K., and Karbhari, V.M., “A Critical Review of Infrared Thermography as a Method for Non-Destructive Evaluation of FRP Rehabilitated Structures,” International Journal of Materials and Product Technology, Vol. 25, No. 4, 2006, 241–266. 15. Mclaughlin, V., Jr., “Defect Detection and Qualification in Laminated Composites by EATF (Passive) Thermography,” Review of Progress in Quantitative Nondestructive Evaluation, Vol. 7B, 1988, pp. 1125–1132. 16. Titman, D.J., “Application of Thermography in Non-destructive Testing of Structures,” NDT & E International, Vol. 34, 2001, pp. 149–154. 17. Bai, W., and Wong, B.S., “Evaluation of Defects in Composite Plates under Convective Environments using Lock-in Thermography,” Measurement Science and Technology, Vol. 12, 2001, pp. 142–150. 18. Rantala, J., Wu, D., and Busse, G., “Lock-in Vibrothermography Applied for Non-destructive Evaluation of Polymer Materials,” Material Science Forum,Vols. 210–213, 1996, pp. 433–438. 19. Darabi, A., and Maldague, X., “Neural Network Based Defect Detection and Depth Estimation in TNDE,” NDT & E International, Vol. 33, 2002, pp.165–175. 20. Abdel-Qader, I., Yohali, S., Abudayyeh, O., and Yehia, S., “Segmentation of Thermal Images for Non-destructive Evaluation of Bridge Decks,”NDT & E International, Vol. 41, No. 5, July 2008, pp. 395–405. 88 21. Sansalone, M., “Impact-Echo: The Complete Story,” ACI Structural Journal, Vol. 94, No. 6, December 1997.Proctor, T., “Some Details of the NBS Conical Transducer,” Journal of Acoustic Emission, Vol. 1, No. 3, 1982, pp. 173–178. 22. Lin, J.-M., Sansalone, M., and Streett, W.B., “Procedure for Determining P-Wave Speed in Concrete for Use in Impact-echo Testing Using a P-Wave Speed Measurement Technique,” Materials Journal, Vol. 94, 1997, pp. 531–539. 23. ASTM C1383-98a, 1998, \"Standard Test Method for Measuring the P-Wave Speed and the Thickness of Concrete Plates Using the Impact-Echo Method,\"ASTM International, West Conshohocken, PA, 1998. 24. Lin, T., Structural Health Monitoring and Its Application to a Bridge with Sprayed Fibre Reinforced Polymer Repair, M.A.Sc. Thesis, The University of British Columbia, Vancouver,, British Columbia, 2007, pp. 89–94. 25. Banthia, N., “Sprayed Fiber Reinforced Polymer: Coming to a Structure near You,” Canadian Civil Engineer, Winter 2005–2006, pp. 8–12. 26. Texas Department of Transportation, Bridge Design Manual, Texas Department of Transportation, Austin, Texas, December 2001. 27. Environment Canada, 2008, http://www.climate.weatheroffice.ec.gc.ca/ [accessed November 2008]. 28. Labossière, P., Rochette, P., Neale, K.W., and Demers, M., 2005. “FRP- Strengthened Structures: Monitoring Issues from Quebec Applications,” Sensing Issues in Civil Structural Health Monitoring, F. Ansari, ed., Springer Netherlands, 2005, pp. 117–126. 29. Khalighi, Y., 2005, “Structural Health Monitoring of Structures Repaired with FRP,” Structural Studies, Repairs and Maintenance of Heritage Architecture IX, 2005, pp. 567–574. . 30. Briukhanov, A.V., Grigorian, S.S., Maigkov, S.M., Plam, M.Ya, Shurora, I.Ya, Eglit, M.E., and Yakimov, Y.L., “On Some New Approaches to the Dynamics of Snow Avalanches,” Physics of Snow and Ice,Proceedings of the International Conference on Low Temperature Science, Hokkaido University, Sappora, Japan, 1966, H. Oura, ed., Vol. 1., Part 2, pp. 1223–1241. 89 31. ASTM C1583-04, 2004, “Standard Test Method for Tensile Strength of Concrete Surfaces and the Bond Strength or Tensile Strength of Concrete Repair and Overlay Materials by Direct Tension (Pull-off Method),” ASTM International, West Conshohocken, PA, 2004. 32. Shen, H.W., and Roper, A.T., Dynamics of Snow Avalanche (With Estimation For Force on a Bridge, Bulletin of the International Association of Scientific Hydrology, XV, 1, 3/1970. 33. Sansalone, M.J., and Streett, W.B., Impact-Echo-Nondestructive Evaluation of Concrete and Masonry, Bullbrier Press, Ithica, New York, 1997. 90 APPENDIX A. FRP APPLICATION PROCEDURE (TYFO SYSTEMS) 1. Allocate two clean and dry 5 gallon buckets and mark them A and B. 2. Mark the inside wall of bucket: • A at 96 mm from the bottom • B at 45 mm from the bottom • B at ¾ bucket height 3. Replace the concrete mixer blade with liquid mixer blade. 4. Pour liquid A and B up to 96 and 45 mm lines in the corresponding buckets. 5. Cut the fiber to the appropriate dimensions. 6. Pour the contents of B into A and mix for 5 min. 7. Take ½ of mixed A+B and pour into B. 8. Apply the content of bucket A as primer to the surface of concrete with a roller. 9. Pour CABOSOIL powder into bucket B up to the ¾ line. 10. Mix the content of bucket B for 6 min. 11. Apply content of bucket B to the concrete surface. 12. Place the fiber on the concrete sample. 13. Remove any air bubble underneath the fiber and apply more of bucket B on top of the fiber. 14. Place the 2nd layer of fiber. 15. Repeat step 13. 91 APPENDIX B. INFRARED THERMOGRAPHY RESULTS B.1 Safe Bridge Figure B.1.1 1A-East, Safe Bridge Figure B.1.2 1B-East, Safe Bridge Figure B.1.3 Girder 2A-East, Safe Bridge Figure B.1.4 Girder 2B-East, Safe Bridge 92 Figure B.1.5 Girder 5A-East, Safe Bridge Figure B.1.6 Girder 5B-East, Safe Bridge Figure B.1.7 Girder 8A-East, Safe Bridge Figure B.1.8 Girder 5B-East, Safe Bridge Figure B.1.9 Girder 8A-East, Safe Bridge Figure B.1.1 Girder 8B-East, Safe Bridge 93 Figure B.1.2 Girder 10A-East, Safe Bridge Figure B.1.3 Girder 10B-East, Safe Bridge Figure B.1.4 Girder 1A-West, Safe Bridge Figure B.1.5 Girder 1B-West, Safe Bridge Figure B.1.6 2A-West, Safe Bridge Figure B.1.7 Girder 2B-West, Safe Bridge 94 Figure B.1.8 Girder 5A-West, Safe Bridge Figure B.1.9 Girder 5B-West, Safe Bridge Figure B.1.19 Girder 8A-West, Safe Bridge Figure B.1.10 Girder 10A-West, Safe Bridge Figure B.1.11 Girder 10B-West, Safe Bridge Figure B.1.12 Girder 1A-Middle, Safe Bridge 95 Figure B.1.13 Girder 1B-Middle, Safe Bridge Figure B.1.14 Girder 2A-Middle, Safe Bridge Figure B.1.15 Girder 2B-Middle, Safe Bridge 96 B-2 St-Étienne Bridge Figure B.2.1 Column1-Exposed, St-Etienne Figure B.2.2 Column1-Inner, St-Etienne Figure B.2.16 Column2-Exposed, St-Etienne Figure B.2.4 Column2-Inner, St-Etienne Figure B.2.5 Column3-Exposed, St-Etienne Figure B.2.6 Column3-Inner, St-Etienne 97 Figure B.2.7 Column4-Exposed, St-Etienne Figure B.2.8 Column4-Inner, St-Etienne 98 APPENDIX C. IMPACT_ECHO TEST RESULTS C.1 St-Etienne Bridge Figure C.1 Column #1- D16 Figure C.2 Column #1- G17 99 Figure C.3 Column #1- D13 Figure C.4 Column #1- F12 Figure C.5 Column #1- B10 100 Figure C.6 Column #1- B1 Figure C.7 Column #1-F3 Figure C.8 Column #1-A4 101 Figure C.9 Column #1- H6 Figure C.10 Column #1- F7 Figure C.11 Column #3- A11 102 Figure C.12 Column #-E10 Figure C.13 Column #3- G11 Figure C.14 Column #3- E15 103 Figure C.15 Column #3-H16 Figure C.16 Column #3- D17 Figure C.17 Column #3- D6 104 Figure C.18 Column #3- A6 Figure C.19 Column #3- F2 C.2 Leslie Street Bridge Figure C.2920 Column #2- D9 105 Figure C.21 Column #2- J9 Figure C.22 Column #2- J6 Figure C.23 Column #2- J10 106 Figure C.24 Column #2- J0 Figure C.25 Column #2- C0 Figure C.26 Column #2- A2 107 Figure C.27 Column #2- C3 Figure C.28 Column #2- J5 "@en ; edm:hasType "Thesis/Dissertation"@en ; vivo:dateIssued "2009-11"@en ; edm:isShownAt "10.14288/1.0063142"@en ; dcterms:language "eng"@en ; ns0:degreeDiscipline "Civil Engineering"@en ; edm:provider "Vancouver : University of British Columbia Library"@en ; dcterms:publisher "University of British Columbia"@en ; dcterms:rights "Attribution-NonCommercial-NoDerivatives 4.0 International"@en ; ns0:rightsURI "http://creativecommons.org/licenses/by-nc-nd/4.0/"@en ; ns0:scholarLevel "Graduate"@en ; dcterms:title "Practicality and capability of NDT in evaluating long-term durability of FRP repairs on concrete bridges"@en ; dcterms:type "Text"@en ; ns0:identifierURI "http://hdl.handle.net/2429/11980"@en .