Mapping Voids, Debonding, Delaminations, Moisture, and Other Defects Behind or Within Tunnel Linings (2013)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

4C h a p t e r 2 Introduction According to data provided by the Federal Highway Adminis- tration, the vast majority of tunnel linings in the United States use cast-in-place (CIP) reinforced concrete, with a significant number using CIP unreinforced concrete, steel/iron liner plate, or shotcrete. In addition, a significant number of tunnels use CIP concrete and a steel/iron liner plate behind the concrete. According to the advisory expert panel for this project, NDT methods are needed to assess the extent of several major problems with tunnel linings: • Water leakage; • Delaminations and spalling of concrete liners due to reinforcing steel corrosion; • Voids behind and within tunnel linings; • Concrete permeability; • Tiles separating from the tunnel liner; • The integrity of steel liners underneath concrete linings; and • Problems with the integrity of ceiling systems and connec- tions to the tunnel lining. The advisory expert panel indicated that NDT should be able to detect any defect within or immediately behind the tunnel linings that have a minimum surface area of 1 sq ft, and any defect needs to be located within 1 ft of the actual location on the tunnel lining. The panel also indicated that NDT should be able to identify delaminated areas and voids up to 4 in. deep as measured from the lining surface with an accuracy within 0.25 in. The advisory expert panel stated that NDT hardware devel- oped for in-depth assessment of tunnel linings should be a simple screening tool—a handheld device—that inspectors can easily use to rapidly detect, locate, and report tunnel lining defects. The panel noted that NDT should facilitate the process of locating and calculating quantities for areas to be repaired. research and Development plan Based on the findings indicated above, the team produced a research and development plan as follows: • An investigation for detecting delaminations, voids, and water intrusion with NDT. The investigation involved con- crete, shotcrete, and steel test specimens constructed at the Texas A&M Transportation Institute (TTI) Riverside Annex. It employed the NDT techniques of ultrasonic tomography, IE, ultrasonic surface waves, air-coupled GPR, ground- coupled GPR, and thermography. And it used 11 concrete and 13 shotcrete specimens; each slab was 6 ft by 6 ft. To simulate delaminations, the team placed plastic sheets in the concrete specimens and thin cloth sheets in the shotcrete specimens. To simulate air voids, the team placed 1-in.-thick Styrofoam wrapped in plastic in the specimens. To simulate water-filled voids, the team placed water-filled plastic bags approximately 1-in. thick in the specimens. The first set of specimens included six intact concrete slabs with thick- nesses of 12 in., 15 in., 18 in., and 24 in., and three defective 15-in.-thick slabs with 1-ft by 1-ft delaminated zones embed- ded in the center of the slabs. The three defective slabs con- tained defects at depths of 1 in., 2 in., and 3 in. from the top surface. Two additional concrete slabs in this set were 15 in. thick with embedded air voids and water voids at a depth of 8 in. The second set of slabs used shotcrete and included four intact slabs with thicknesses of 4 in., 6 in., 8 in., and 12 in., and five 12-in.-thick delaminated slabs. The 1-ft by 1-ft delaminated areas were embedded at the center of each slab at depths of 1 in., 2 in., 3 in., 4 in., and 8 in. from the top sur- face. Four additional shotcrete slabs contained air voids and water voids of different sizes at different depths. The team also used specimens containing clay lumps constructed under another TTI study, a concrete bridge deck constructed by the University of Texas at El Paso (UTEP) for another SHRP 2 study, a continuously reinforced concrete pavement Research Approach

5section on I-20 in Fort Worth, and an airport runway section at the George Bush Intercontinental Airport. • Field validation testing of NDT devices using actual tunnels. A pilot project for the SPACETEC equipment was conducted for the Chesapeake Channel Tunnel in April 2011. In addi- tion, initial tests with air-coupled GPR and thermal cameras were conducted using two tunnels in Helsinki, Finland. Finally, the team conducted tunnel testing in Colorado, Texas, and Virginia. • An investigation for detecting loose tiles and moisture underneath tiles using NDT. The NDT techniques used in this investigation are air-coupled GPR, thermal cameras, and sounding. The team is using a tiled surface in an actual tunnel for this ongoing investigation. • Development of NDT for measuring concrete permeability. This step involved a laboratory study to correlate NDT mea- surements with concrete specimens that have different per- meabilities, and field verification using existing concrete tunnel linings. The NDT techniques used in this investiga- tion were the dielectric probe, air-coupled GPR, resistivity, and ultrasonic surface waves. As described in this report, concrete permeability cannot be measured directly in the field using air-coupled GPR; however, the team did generate recommendations that relate potential for corrosion to GPR dielectric measurements. In addition, the report indicates how permeability could be estimated if future NDT can measure certain properties. test Specimens Concrete and Shotcrete Specimens with Simulated Delaminations and Voids Eleven normal-weight concrete slabs and 13 shotcrete slabs were constructed to simulate various defects. The concrete slabs were used to mimic typical concrete tunnel linings with and without reinforcing steel. The shotcrete slabs were con- structed to mimic applications in which shotcrete is sprayed on as a finished layer, as typically found in tunnel linings. A specially designed lattice girder, also typical in tunnel wall construction, was used as reinforcement in the shotcrete slabs (Figure 2.1, bottom right). The simulated delaminations in the slabs were constructed from three types of material. Delaminations were imitated by using 0.05-mm (0.002-in.) plastic square sheets and 0.25-mm (0.01-in.) cloth squares (Figure 2.1, top right). Air-filled voids (Figure 2.1, top left) were constructed by inserting foam squares 13 mm (0.5 in.) thick in vacuum-sealed plastic bags. Water-filled voids (Figure 2.1, bottom left) were constructed in a similar manner by placing water-filled Ziploc bags within vacuum-sealed plastic bags and carefully padding the defect Figure 2.1. Construction of slabs with simulated defects.

6with concrete/shotcrete during construction so as not to puncture the plastic. Table 2.1 provides a summary of the specimen details. Concrete Specimens with Simulated Clay Lumps In addition to the concrete and shotcrete slabs, six concrete slabs were tested that TTI had constructed in the 1990s as part of a previous research project. These slabs contain manufactured clay lumps of different diameters. The clay lumps are a high- plasticity clay, classified as Burleson Clay CH (AASHTO A-7-6) with a plasticity index (PI) range of 35 to 45. The slabs and lumps are shown in Figure 2.2 and are summarized in Table 2.2 (Specimens A2 through F2). These six specimens consist of two sets of three slabs: one set with steel reinforcement and one set without. In each set, one slab was designated as the control with no clay lump contaminations. The remaining two had various Table 2.1. Summary of Concrete and Shotcrete Slab Specimens with Simulated Defects Specimen Name Specimen Depth (mm) Material Reinforced Detail Defect True Depth of Defect (mm) Alpha 305 Concrete None None NA Beta 457 Concrete d = 127 mma Natural crack NA Gamma 305 Concrete d = 127 mma None NA Delta 610 Concrete None None NA Epsilon 610 Concrete d = 127 mma None NA Zeta 381 Concrete d = 127 mma None NA Eta 381 Concrete d = 127 mma 0.05-mm thin plastic 51 from top Theta 381 Concrete d = 127 mma 0.05-mm thin plastic 76 from top Iota 381 Concrete d = 127 mma 0.05-mm thin plastic 25 from top Kappa 381 Concrete d = 127 mma Air-filled void (13-mm Styrofoam) 203 from top Lambda 381 Concrete d = 127 mma Water-filled void (Ziploc bag) 203 from top A 102 Shotcrete None None NA B 152 Shotcrete None None NA C 203 Shotcrete None None NA D 305 Shotcrete One lattice girder in center of slab, sitting on bottom form Air-filled void (13-mm Styrofoam) 193 from top E 305 Shotcrete One lattice girder in center of slab, sitting on bottom form Water-filled void (Ziploc bag) 191 from top F 305 Shotcrete One lattice girder in center of slab, sitting on bottom form Air-filled void (13-mm Styrofoam) 76 from top G 305 Shotcrete One lattice girder in center of slab, sitting on bottom form Water-filled void (Ziploc bag) 76 from top H 305 Shotcrete One lattice girder in center of slab, sitting on bottom form 0.25-mm thin cloth 203 from top I 305 Shotcrete One lattice girder in center of slab, sitting on bottom form 0.25-mm thin cloth 102 from top J 305 Shotcrete One lattice girder in center of slab, sitting on bottom form 0.25-mm thin cloth 76 from top K 305 Shotcrete One lattice girder in center of slab, sitting on bottom form 0.25-mm thin cloth 51 from top L 305 Shotcrete One lattice girder in center of slab, sitting on bottom form 0.25-mm thin cloth 25 from top M 305 Shotcrete One lattice girder in center of slab, sitting on bottom form None NA Note: All slab specimens are nominally 1.83 m by 1.83 m. NA = not available. a Two mats of No. 5 rebar, at depth d from top and bottom, 203 mm on center.

7levels of lumps of documented sizes corresponding to three regions of interest: (1) lumps below the reinforcement that rep- resent typical lumps dense enough not to be quickly displaced toward the surface via vibration, (2) lumps that are caught in the reinforcing steel layer on their path toward the surface, and (3) lumps that are dispersed between the reinforcement and the top surface. The depth of the slabs is nominally 305 mm (12 in.), but all measurements are taken as approximate since neither ground truth data were retrieved nor any accurate pictures were taken to confidently support documented placement. Concrete Bridge Deck with Simulated Defects In addition to the above-mentioned slabs, a bridge deck con- structed by UTEP was available for blind testing. The bridge deck was constructed with known artificial delaminations, cracks, and corroded reinforcement. Several parameters were considered in the construction of the artificial delaminations. These included stacked delaminations and delaminations of various thicknesses, ranging from 0.3 mm (0.01 in.) to 2.0 mm (0.08 in.); sizes ranging from 305 mm by 305 mm to 610 mm by 1220 mm (12 in. by 12 in. to 24 in. by 48 in.); and depths above reinforcing steel at 64 mm (2.5 in.) below the surface and below two layers of reinforcing steel at 152 mm (6 in.), with some located above prestressed girders supporting the slab. The deck, as shown in Figures 2.3 and 2.4, measures 2.4 m by 6.1 m by 0.2 m (8 ft by 20 ft by 8.75 in.) and rests on three pre- stressed concrete girders. Simulated defects constructed in the deck consist of nine artificial delaminations, five cracks, and two corroded reinforcement mats, all of which are summarized in Table 2.3. In constructing the deck, 27.6 MPa (4,000 psi) concrete was used, and two layers of No. 5 longitudinal and transverse steel were placed at 254 mm and 203 mm (10 in. and 8 in.) on center, respectively, at centroid depths of 83 mm and 184 mm (3.25 in. and 7.25 in.) from the surface. The 28-day strength and modu- lus exceeded 34.5 MPa (5,000 psi) and 27.6 MPa (4,000 psi), respectively. A 0.25-mm (0.01-in.) polyester fabric was used to mimic an ultrathin horizontal delamination. The vertical cracks were constructed from both thick and thin cardboard Figure 2.2. Clay lump slab construction. Table 2.2. Summary of Concrete Specimens with Simulated Clay Lumps Specimen Name Specimen Depth (mm) Material Reinforced Detail (mm) Defect True Depth of Defect (mm) A2 305 Concrete d = 152a None NA B2 305 Concrete d = 152a Large (152-mm ∅) clay lumps 152 from top C2 305 Concrete d = 152a Medium (102-mm ∅) clay lumps 76, 152, and 229 from top D2 305 Concrete d = 152a None NA E2 305 Concrete d = 152a Large (152-mm ∅) clay lumps 152 from top F2 305 Concrete d = 152a Medium (102-mm ∅) clay lumps 76, 152, and 229 from top Note: NA = not available. a Two mats of No. 5 rebar, at depth d from top and bottom, 203 mm on center.

8sheets. The No. 5 corroded steel mats were electrically merged and attached to the normal reinforcement. The corrosion depth was measured to be 1–2 mm (0.04–0.08 in.) before pouring the concrete. tunnels tested in the Study Chesapeake Channel Tunnel, Virginia The Chesapeake Channel Tunnel (Figure 2.5) is one of two tunnels that make up the Chesapeake Bay Bridge-Tunnel sys- tem, joining southeastern Virginia to the Delmarva Peninsula. Hailed worldwide as a modern engineering wonder, the 37-km (23-mi) long system includes 3.2 km (2 mi) of causeway, four manmade islands, 8.9 km (5.5 mi) of approach roads, 19.3 km (12 mi) of low-level trestle, two 1.6-km (1-mi) steel tunnels, and two bridges. The Chesapeake Channel Tunnel (during construction and briefly afterward it was called the Baltimore Channel Tunnel) was constructed using a cut-and-cover method. Precast steel tubes, fabricated and assembled in Orange, Texas, were floated to a shipyard in Norfolk, Virginia, where the reinforced concrete linings and roadway were con- structed. The sections were floated to the site before being sunk into a trench. Each steel tube, 90 m (300 ft) in length and 11 m (37 ft) in diameter, was joined to the other, sealed, and connected to its adjoining section. As each steel section was welded together, patches between the 90-m (300-ft) sections had to be formed with concrete to make an overlapping seal. Eisenhower Memorial Tunnel, Colorado The Eisenhower Memorial Tunnel, located approximately 97 km (60 mi) west of Denver, Colorado, is one segment of a 2.7-km (1.7-mi) dual bore project started in 1968. Shown in Figure 2.6, the Eisenhower Memorial Tunnel, which carries I-70 west, is paired with the Edwin C. Johnson Memorial Tunnel, which carries eastbound I-70. Although the east- bound bore was not completed until almost 1980, construc- tion on the Eisenhower bore was completed by 1973. Built using drill and blast methods through a mountain with a max- imum overburden of 448 m (1,470 ft), the average tunnel dimensions are 14.6 m in height (48 ft) and 12.2 m (40 ft) in width. In 2011, the average daily traffic was 28,155 vehicles. All areas of interest evaluated within the tunnel were tested from inside the plenum (above the traffic, Figure 2.7). Hanging Lake Tunnel, Colorado Completed in 1992 with a maximum length of 1,219 m (4,000 ft) through the southern wall of Glenwood Canyon, Hanging Lake Tunnel (Figure 2.8) was the last link in the Figure 2.3. Simulated bridge deck at UTEP in El Paso, Texas. Figure 2.4. Layout of constructed bridge deck.

9Table 2.3. Summary of Simulated Defects in Concrete Bridge Deck Simulated Defect Defect Material Actual Dimension (mm) Actual Depth (mm) Delamination (DL 1) Soft, high-strength 1-mm Styrofoam 305 by 305 64 Delamination (DL 2) Soft, high-strength 1-mm Styrofoam 610 by 610 64 Delamination (DL 3) Soft, high-strength 1-mm Styrofoam 610 by 610 64 Delamination (DL 4) Soft, high-strength 2-mm Styrofoam 305 by 305 64 Delamination (DL 5) Soft, high-strength 2-mm Styrofoam 610 by 610 64 Delamination (DL 6) Soft, high-strength 2-mm Styrofoam 610 by 610 64 Delamination (DL 7) Soft, high-strength 1-mm Styrofoam 610 by 610 152 Delamination (DL 8) Soft, high-strength 1-mm Styrofoam 610 by 1,219 152 Delamination (DL 9) Soft, 0.25-mm polyester fabric 305 by 610 64 Vertical crack (CK 1) Soft, thin cardboard 305 long 64 Vertical crack (CK 2) Soft, thin cardboard 305 long 64 Vertical crack (CK 3) Soft, thick cardboard 305 long 76 Vertical crack (CK 4) Soft, thick cardboard 305 long 152 Vertical crack (CK 5) Natural crack (observed after construction) 330 long 64 Corroded reinforcement (CR 1) 1–2 mm deep corrosion, No. 5 bars 762 by 762 76 Corroded reinforcement (CR 2) 1–2 mm deep corrosion, No. 5 bars 762 by 762 165 Figure 2.5. Chesapeake Channel Tunnel: entrance (left) and interior view (right).

10 Interstate highway system. Both bores of the tunnel were built using multiple-face drill and blast methods. Between the westbound and eastbound bores, a four-story control center monitors traffic along I-70, fully equipped with emergency response vehicles and trained staff. No Name Tunnel, Colorado The No Name Tunnel was constructed in 1965 and is located approximately 7.5 miles west of the Hanging Lake Tunnel. The team collected air-coupled GPR data and infra- red images only in the westbound bore, which is approxi- mately 1,000 ft long. The upper portion of the tunnel has a concrete surface; the sides are tiled. Figure 2.9 shows the TTI Figure 2.6. Eisenhower Memorial Tunnel, Colorado. Figure 2.7. Eisenhower Memorial Tunnel plenum view. Figure 2.8. Hanging Lake Tunnel: exterior (left) and interior plenum view (right).

11 air-coupled GPR system collecting data in the tunnel. Plan sets were not available. Washburn Tunnel, Texas The Washburn Tunnel (Figure 2.10), the only underwater vehi- cle tunnel in operation in Texas, was completed in 1950 and car- ries a federal road beneath the Houston Ship Channel, joining two Houston suburbs. The tunnel was constructed using the immersed-tube method, with sections joined together in a pre- pared trench, 26 m (85 ft) below water. The entire inner wall is tiled with 110-mm by 110-mm (4.3-in. by 4.3-in.) ceramic tiles. NDt Devices and techniques Used in the Study Air-Coupled Ground-Penetrating Radar Ground-penetrating radar sends discrete electromagnetic pulses into a structure and then captures the reflections from layer interfaces in the structure. Radar is an electromagnetic wave and therefore obeys the laws governing reflection and transmission of electromagnetic waves in layered media. At each interface within a structure, part of the incident energy is reflected, and part is transmitted. The amplitude of radar reflections and the time delay between reflections are used to calculate layer thicknesses and layer dielectrics. For purposes of this study, the surface layer dielectric is of most interest. This value is calculated as follows: ( ) ( )ε = + −       1 1 1 1 2A A A A a m m where ea = dielectric of lining surface, A1 = amplitude of reflection from surface in volts, and Am = amplitude of reflection from a large metal plate in volts (this represents the 100% reflection case). Because air-coupled systems (Figure 2.11) are not in con- tact with the structure, data collection can theoretically hap- pen at full traffic speeds, although this is not practical for tunnel lining data collection. Air-coupled antenna systems are manufactured by GSSI, Penetradar, Pulse Radar, and Wavebounce—all from the United States. Butterfly dipole systems are manufactured by Radarteam Sweden AB. The system used by TTI to collect tunnel lining data includes a Wavebounce 1-GHz central frequency GPR antenna with distance measuring indicator equipment. The system uses data collection software developed by TTI. Researchers used the Pavecheck and Colormap programs, also developed by TTI, to analyze the data. The researchers slightly modified Pavecheck and renamed it Tunnelcheck. (This software is available for free Figure 2.9. TTI air-coupled GPR system collecting data in No Name Tunnel. Figure 2.10. Washburn Tunnel: entrance ( left) and interior view (right).

12 download; the user’s manual is provided in a separate publica- tion.) The researchers also mounted a FLIR T300 camera on the GPR boom and the TTI data collection system collected images from this camera along with the GPR data. The penetration depth of air-coupled GPR is usually around 24 in. for a 1-GHz system. Although air-coupled GPR can detect reinforcing steel, it cannot detect defects in concrete unless the defects contain significant air pockets or significant moisture. Nonetheless, the research team believes that the surface dielec- tric can be used to determine where to conduct testing with in-depth NDT devices and techniques. Appendix A contains the air-coupled GPR testing criteria. Appendix K contains data analysis results from the air-coupled GPR tunnel and specimen testing conducted under this study. Ground-Coupled Ground-Penetrating Radar Ground-coupled GPR needs to be either in contact with or close to the lining surface when collecting data (Figure 2.12). The operating principles are the same as air-coupled GPR: ground-coupled GPR cannot detect defects in concrete unless the defects contain significant air pockets or significant mois- ture. However, ground-coupled GPR can detect defects that air-coupled GPR cannot. Ground-coupled GPR can also detect reinforcing steel. Researchers used the GSSI 1.5-GHz central frequency GPR antenna during the tunnel tests because shal- low defects were found during those tests; researchers used a 900-MHz central frequency GPR antenna during the TTI spec- imen tests because the researchers were trying to determine if ground-coupled GPR can detect deep defects. Appendix B contains the ground-coupled GPR testing criteria. Appendix Q contains data analysis results from the ground-coupled GPR testing conducted under this study. Handheld Thermal Camera Handheld thermal cameras (Figure 2.13) have improved sig- nificantly over the past decade, with consistently higher image resolutions and improved temperature accuracy occurring over time. The research team used the FLIR T300 thermal camera. The researchers analyzed the images for changes in tunnel lining temperature, which could indicate possible defects within or behind the lining. The team believes that the images from such cameras can be used to determine where to conduct testing with in-depth NDT devices and techniques. Appendix C contains the handheld infrared camera testing criteria. Appendix L contains selected images from tunnels and TTI test specimens. Ultrasonic Tomography The ultrasonic tomography (UST) system used in this study is a device with an array of ultrasonic transducers that transmit and receive acoustic stress waves for the inspection of concrete structures. The system used here, the A1040 MIRA, is produced by Acoustic Control Systems (Figure 2.14). Figure 2.11. TTI air-coupled GPR system collecting tunnel roof lining data in Colorado. Figure 2.12. Ground-coupled GPR equipment. Figure 2.13. FLIR T300 thermal camera used in the study.

13 The tomograph, shown in Figure 2.14 (left), uses a 4-by-12 grid of mechanically isolated and dampened transducers that can fit the profile of a rough concrete testing surface with a vari- ance of approximately 10 mm (0.4 in.). Each row of four trans- ducers transmits stress waves sequentially while the remaining rows act as receivers. In this manner, there is a wide coverage of shear wave pulses that reflect at internal interfaces where the material impedance changes. With the help of a digitally focused algorithm, a three- dimensional (3-D) volume is presented with each point of possible reflection in half-space represented by a color scheme, scaled according to reflecting power. This 3-D image can also be dissected into each of the three planes representing its vol- ume: the B-scan, C-scan, and D-scan (Figure 2.15). The B-scan is an image slice showing the depth of the specimen on the vertical (or z) axis versus the width of scan on the horizontal (or x) axis. This slice is a plane perpendicular to the scan- ning surface and parallel to the length of the device. The C-scan is an image slice showing the plan view of the tested area, with the vertical (or y) axis of the scan depicting the width parallel to the scanning direction and the horizontal (or x) axis of the scan representing the length perpendicular to the scanning direction. Note that the scanning direction is always defined as the y-axis as seen in Figure 2.15. The D-scan is like the B-scan in that it images a plane perpen- dicular to the testing surface, but it is oriented parallel to the scanning direction. On each of the scans, the various inten- sities reported by the returned waves are color-coded from light blue to deep red, representing low reflectivity (typically sound concrete) and high reflectivity (any type of imped- ance), respectively. With this intensity scaling, any disconti- nuities are readily apparent, with distinctly different wave speeds such as voids, delaminations, cracks, and other abnor- malities. This UST system has had limited exposure to indus- trial applications but is quickly becoming recognized as a powerful NDT method. Appendix D contains the ultrasonic tomography testing criteria. Appendices M and N contain the testing results from tunnel linings and test specimens, respectively. Ultrasonic Echo An ultrasonic transducer is used to generate and/or receive ultrasonic waves in/from a test medium. Ultrasonic echo tech- nique involves sending and receiving ultrasonic pulses from the same side of the test object, by the same or two separate transducers. The ultrasonic pulse velocity is correlated to material strength or quality. The measurement of propaga- tion time is used to localize cracks, voids, and delamination and/or to estimate the thickness of a structure. Structural boundaries and defects that are large enough (with respect to the ultrasonic wavelength) induce a high contrast in acoustic impedance and result in the reflection of ultrasonic waves. The reflected waves are detected in ultrasonic scans, and the two-way travel time is used to estimate the reflector location Figure 2.14. A1040 MIRA system ( left) and transmission/reception of acoustic waves and corresponding echo intensity (right). Figure 2.15. B-scan, C-scan, and D-scan relative to the tomograph.

14 (assuming or knowing the ultrasonic wave velocity in the test medium). The handheld ultrasonic transducer used by the Federal Institute for Materials Research and Testing (BAM) in Ger- many for field testing together with the corresponding data acquisition/analyzer unit is shown in Figure 2.16. In tunnel testing applications, the ultrasonic echo technique can be used to estimate the thickness of the tunnel lining and to detect delamination and voids within the lining. Appendix E contains the ultrasonic echo test criteria and Appendix Q contains data analysis results from the ultrasonic echo testing conducted under this study. Ultrasonic Surface Waves and Impact Echo Methods with the Portable Seismic Property Analyzer Ultrasonic Surface Wave Method The ultrasonic surface wave (USW) method is used to esti- mate the average velocity of propagation of surface waves in a medium, based on the time at which different types of energy arrive at each sensor (Figure 2.17a). The velocity of propagation, VR, is typically determined by dividing the dis- tance between two receivers, DX, by the difference in the arrival time of a specific wave, Dt. Knowing the wave velocity, E, the modulus can be determined from shear modulus, G, through Poisson’s ratio (ν) by using 2 1E G( )= + ν Shear modulus can be determined from shear wave velocity, VS, by using = γ 2G g VS Figure 2.16. Ultrasonic echo equipment A1220 Monolith by ACSYS. h Intact Intact Severe Delamination Intact Flexural Mode (a) USW method (b) Impact echo method Source: Gucunski and Maher 1998 Figure 2.17. Schematic illustration of test methods: (a) ultrasonic surface wave method and (b) IE method.

15 The modulus from surface wave velocity, VR, first converted to shear wave velocity, can be determined by using ( )= − ν1.13 0.16V VS R In the USW method, the variation in velocity with wave- length is measured to generate a dispersion curve. For a uni- form or intact tunnel lining, the dispersion curve shows more or less a constant velocity within the wavelengths no greater than the thickness of the slab. When a delamination or void is present in a concrete slab or the concrete has deteriorated, the average surface wave velocity (or modulus) becomes less than the actual modulus because of interference from the defect. In this case, the velocity or modulus obtained may be called an apparent velocity or modulus. Impact Echo Method The IE method is one of the most commonly used NDT methods for detecting delamination in concrete. This method works by striking a plate-like object such as a tunnel lining with an impactor that generates stress waves at frequencies up to 20 kHz to 30 kHz and collecting signals with a receiver (see Figure 2.17b). By using a fast Fourier transform (FFT) algo- rithm, the recorded time domain signal is converted into a frequency domain function (amplitude spectrum), and the peak frequency is monitored. For an intact point on a slab, the thickness (h) is then determined from the compression wave velocity (Vp) and the return frequency ( f ): 2 h V f p = α where a is about 0.96 for concrete slabs. For a deep and relatively small delaminated location in a tunnel lining, the return frequency may shift to a higher fre- quency corresponding to the depth of the delamination. As shown in Figure 2.17b, a shallow or a deep but extensive and severely delaminated area is usually manifested by a low peak frequency, indicating that little or no energy propagates toward the bottom of the deck, and a flexural mode dominates the frequency response. In this case, the equation is not applicable to measure the depth of delamination since it is influenced by several factors. Description of the PSPA USW and IE measurements can be performed with these two methods simultaneously with the PSPA shown in Figure 2.18. The traditional PSPA is a box containing a solenoid-type impact hammer and two high-frequency accelerometers (Fig- ure 2.18a). All controls and data acquisition are in a computer connected to the box. The two receivers allow the calculation Vp using the USW method. The test at a single point is simple and takes less than 30 s. The impact duration (contact time) is about 60 µs, and the data acquisition system has a sampling frequency of 390 kHz. As shown in Figure 2.18b, the PSPA has been redesigned to make it more user friendly and compact for tunnel work. The new PSPA is self-contained and eliminates the need for an external computer to collect data. The waveforms collected in the field are stored in a removable flash memory. The new PSPA is also lighter compared with the traditional PSPA (8 lb versus 16 lb). Data collection with the new PSPA is a two- hand operation, which can accommodate the curvature within the tunnel more easily. Data acquisition with the new PSPA is on average two to three times faster than with the traditional one. The new PSPA is also equipped with three receivers to better optimize the data collection for the com- bined IE-USW methods. The power source for the device is six AAA batteries placed in a container that operators can carry on their belts. Typical signals collected with the PSPA are shown in Figure 2.19. These signals are used to develop USW dispersion curves and the IE amplitude spectra. The advantage of combining USW and IE methods in a single device is that once the test is performed, the variations in the modulus (an indication of the quality of concrete) and return resonance frequency (an indication of the full thickness or depth of delamination) of a slab can be assessed concurrently. (b) New version (a) Traditional device Figure 2.18. Portable seismic property analyzer (PSPA).

16 Figure 2.20 compares typical USW dispersion curves from an intact area and a defective area. The dispersion curve shifts to lower moduli in defective areas. The amplitude spectra for typi- cal intact and defective points are shown in Figure 2.21. Based on an average compression wave velocity of about 14,000 ft/s measured for the concrete, the dominant frequency corre- sponding to the tunnel thickness (15 in.) is around 5.4 kHz. Compared with the intact point, higher peak frequencies mostly control the response at the defective points. SPACETEC Scanner The SPACETEC scanner (Figure 2.22) is a mature system devel- oped specifically for the inspection of railway and roadway tunnels. Therefore, employing this technology for this proj- ect required no additional hardware and software develop- ment. The scanner system has been used to survey many miles Figure 2.19. PSPA sample test results. 2 3 4 5 6 7 8 9 10 11 12 2000 2500 3000 3500 4000 4500 5000 5500 D ep th (in .) Modulus (ksi) Defective Point Intact Point A ve ra ge M od ul us = 4 52 7 ks i A ve ra ge M od ul us = 2 62 7 ks i Figure 2.20. Typical dispersion curves for defective and intact points. 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 x 104 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 X: 8387 Y: 1 Frequency (Hz) N or m a liz e d A m pl itu de X: 5337 Y: 1 Intact Point Defective Point Figure 2.21. Typical amplitude spectra for intact and defective points. Appendix F contains the ultrasonic surface waves and impact echo testing criteria.

17 of railway and roadway tunnels (mostly railway) in various countries but was never used in the United States before this project. The SPACETEC TS3 scanner records three different mea- surements in a single pass: • Survey of the cross-sectional tunnel profile; • Full-surface visual recording of the tunnel lining; and • Full-surface thermographic recording (thermal imaging) of the tunnel surface. The scanner processes the measurements, which can then be viewed individually or together to detect and locate tunnel sur- face and near-surface anomalies. The high-resolution visual recording allows a thorough inspection of the tunnel surface and, combined with the profiling, the location of surface defects. The cold spots in the thermal images are usually indications of near-surface moisture. Superimposing the thermal images on the visual recordings allows such moist zones to be easily identified. Monitoring changes in the tunnel profile over time presents another potential application of this system. Appendix I contains results of the SPACETEC testing in the Chesapeake Channel Tunnel. Other NDT Devices and Techniques Used in the Study The research team used resistivity and dielectric probe devices during this research. However, the devices were only useful in a laboratory environment and are not recommended for use in tunnel lining field tests. Appendix S contains the results of a laboratory study that attempted to correlate dielectric (or per- mittivity) measurements to concrete permeability. Researchers used resistivity and dielectric probe equipment during this laboratory study. Researchers also attempted to develop an acoustic sounding technique for detecting delaminated tiles. Appendix O con- tains a description of the technique and the results obtained so far. This technique is still under development; thus, it is not ready for implementation. The team did collect thermal data in Finnish and U.S. tun- nels with the FLIR A325 vehicle-mounted thermal camera. This camera has the same thermal measurement specifications as the FLIR T300 handheld thermal camera. Roadscanners developed commercial software before this SHRP 2 study began that collects and helps analyze such data for the FLIR A325 camera. Although the results from the testing are promising, the team does not recommend implementation of the system at this time. Further software refinements are needed before this system can be implemented effectively. Appendix H con- tains the system’s testing criteria and Appendix L contains images from the system. As described later in this report, the research work in Finland also involved the use of laser scanning systems. Although the data analysis results and images from those systems did not apply directly to the goals of this project, the testing results in Finland indicated that laser scanning systems provide inter- esting and useful data relating to the shape (or profile) and the surface condition of tunnel linings. Appendix J offers more information about the results of testing with these systems. Dr. Fulvio Tonon, one of the authors, conducted digital photogrammetry work in three tunnels during the course of this project. Although the data analysis results and images from this technique did not apply directly to the project, the results may be of interest to the reader. Appendix X contains a description of, and results from, this technique. Figure 2.22. SPACETEC scanner in the Chesapeake Channel Tunnel.