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.

238 a p p e N D I x Q testing program The Federal Institute for Materials Research and Testing (BAM) in Germany carried out field testing between October 3 and 12, 2011, in three tunnels in the United States: two in Colorado (Eisenhower Memorial Tunnel and Hanging Lake Tunnel) and one in Virginia (Chesapeake Bay Bridge-Tunnel). In each tunnel, selected areas were tested using three nondestructive testing (NDT) techniques: ground-penetrating radar (GPR), ultrasonic echo (designated as US in test area sketch figures), and impact echo (IE). The allocated testing time in each tunnel was limited. The number and location of the test areas were selected on the basis of either preanalysis (mostly thermo g- raphy) or the existence of visual distress. The on-site work- ing conditions were also taken into account. Table Q.1 provides the details of the test program, including the number and size of test areas in each tunnel as well as the testing methods. tunnel Description Eisenhower Memorial Tunnel The Eisenhower Memorial Tunnel is located approximately 60 mi west of Denver, Colorado, on I-70 and is a part of the Colorado Department of Transportation. It is the highest vehicular tunnel in the world, located, on average, at an eleva- tion of 11,112 ft. It is 1.693 mi long and runs through a moun- tain within the Arapaho National Forest. Figure Q.1 shows a construction information bulletin from the tunnel side. Fig- ure Q.2 shows the entrance to the tunnel as well as one of the supply air ducts where measurements took place. Construction started in March 1968 and was completed in March 1973. The information about this tunnel was obtained from the Colorado Department of Transportation website (Colorado DOT 2011). Hanging Lake Tunnel The Hanging Lake Tunnel stretches more than 4,000 ft through a mountain bordering the Glenwood Canyon in Colorado as part of I-70. The most noteworthy feature of the tunnel is the direct change of I-70 from bridge to tunnel (Figure Q.3). Construction started in 1980 and was completed in 1992. The entrance to the tunnel and the supply air duct are shown in Figure Q.3. Chesapeake Bay Bridge-Tunnel The Chesapeake Bay Bridge-Tunnel is part of a bridge-tunnel system connecting Virginia’s Eastern Shore with its mainland. The tunnel itself is 1 mi long, going under the Atlantic Ocean. Construction started in November 1960, and the first part was opened in April 1964. Figure Q.4 shows a bulletin board from the side and a view of the supply air duct where most measurements were taken. Measurement techniques The different measurement equipment BAM used for this proj- ect can be mounted on an automated scanning device devel- oped by BAM. The ZFP scanner (Figure Q.5) (Zoega et al. 2012) can be used on horizontal surfaces as well as vertical surfaces, including overhead testing, even in narrow areas. The scanner is fixed to the surface using vacuum “feet” or plates. When run- ning acoustic tests requiring contact, choosing 1-in.-grid spac- ing allows a speed of operation of 11 sq ft/h (1 sq m/h). With noncontact transducers such as air-coupled radar antenna, test- ing at a speed of 156 sq ft/h would be possible. The field of mea- surement can be up to 17.6 sq ft (4.2 ft by 4.2 ft). The advantage of the ZFP scanner is its easy and fast on-site assembly. It can be carried in a relatively light and small pack- age. Its size allows the scanner to be transported in cars and carried through small openings to reach difficult-to-access areas such as the vents above tunnels. The commonly used equipment for NDT of structures—including GPR, ultrasonic echo, and IE devices—can be easily attached to the scanner for testing and detached after completing the measurements. The scanning and NDT data acquisition are controlled by a single BAM Testing in U.S. Tunnels

239 Table Q.1. Overview of BAM Field-Testing Program, October 2011 Code Tunnel Location Date Technique Size (in.) Spacing (in.) Notes EH1 Eisenhower Seg 3 Block 2-3 10/03/2011 US 40 × 40 1 IE 2 × 24 1 EH2 Eisenhower Seg 11 Block 1 10/04/2011 US 40 × 24 1 anomaly, reinforcement IE 40 × 24 1 GPR 40 × 24 2 anomaly, reinforcement HL1 Hanging Lake Seg 56/57 10/05/2011 US 48 × 24 1 anomaly IE 48 × 24 1 anomaly GPR 48 × 24 2 beam, dowels, reinforce HL2 Hanging Lake Seg 57 10/05/2011 US 48 × 24 1 crack, reinforcement IE 48 × 24 1 backwall? GPR 48 × 24 2 crack, reinforcement HL3 Hanging Lake Seg 57/58 10/06/2011 US 48 × 24 1 reinforcement IE 48 × 24 1 GPR 48 × 24 2 joint, dowels CPB1 Chesapeake Bay Bridge St.No. 474+27 10/11/2011 US 48 × 24 1 anomaly, backwall, reinforce. IE 46 × 24 1 anomaly, backwall GPR 48 × 24 2 reinforcement CPB2 Chesapeake Bay St.No. 481-76 10/12/2011 US 40 × 24 1 backwall, anomaly, reinforce. IE 40 × 24 1 backwall, anomaly GPR 40 × 24 2 reinforcement CPB3 Chesapeake Bay St.No. 486-67 10/12/2011 US 48 × 24 1 anomaly, backwall, reinforce. IE 48 × 24 1 backwall, anomaly GPR 48 × 24 2 reinforcement CPB4 Chesapeake Bay St.No. 487 10/12/2011 US 48 × 36 1 IE 4 × 36 1 condition of tile bonding GPR 48 × 36 2 reinforcement Source: Colorado DOT 2012 Figure Q.1. Construction information, Eisenhower Memorial Tunnel.

240 Figure Q.2. Entrance to the Eisenhower Memorial Tunnel (a) and the interior of the supply air duct where measurements took place (b). (a) (b) Figure Q.3. Hanging Lake Tunnel entrance (a) (Salek 2002) and supply air duct where measurements took place (b). (a) (b) Figure Q.4. Construction information for the Chesapeake Bay Bridge-Tunnel (a) and supply air duct where most measurements took place (b). (a) (b)

241 notebook. This simplifies the control and reduces the equip- ment and weight of the measurement system. The three NDT techniques and the typical data from each are briefly discussed below. GPR Basic Operation Principles Ground-penetrating radar is a widely used subsurface scan- ning tool that was employed here to detect subsurface defects in tunnel linings. GPR sends discrete electromagnetic pulses into the structure and then captures the reflections from sub- surface layer interfaces. Radar is an electromagnetic wave and therefore obeys the laws governing reflection and transmis- sion of electromagnetic waves in layered media. At each inter- face within a structure, a part of the incident energy is reflected and a part is transmitted. The ratio of reflected to transmitted energy depends on the electromagnetic contrast of the mate- rial on either side of the interface. Two main types of GPR equipment are typically used for civil structure investigations. High-speed air-coupled GPR systems are capable of testing at speeds up to about 50 mph and can penetrate up to 24 in. in some materials. They are excel- lent tools for network-level data collection. High-frequency ground-coupled GPR systems provide better depth pene- tration and high densities of readings and are excellent for project-level data collection and applications concerned with locating steel and defects such as voids in concrete. Their limitation is that they must stay in close contact with the material being tested, making the speed of data collection relatively slow (1 mph to 5 mph). GPR antennas can emit electromagnetic pulses of different frequencies. The choice of frequency depends on the required depth of penetration and depth resolution. In general, lower- frequency antennae have a better resolution in deeper depth. Higher-frequency antennae show better details of reflectors close to the surface but do not penetrate the test object as deeply. Determining which antenna to use therefore depends on the task, the experience of the user, and other NDT meth- ods used at the scene. In this study, BAM used a ground-coupled GPR from Geo- physical Survey Systems Inc. (GSSI), with a center frequency of about 1.5 GHz (Figure Q.6). Typical Results Ground-coupled GPR has proved very useful in discovering reinforcements, dowels, surface cracks, moisture, and other intrusions. As shown in Figure Q.7, scanner testing on fine grids provides the opportunity to generate B-, C-, and D-scans from the measurements. Reinforcement bars and dowels are best seen in C-scans and recognized by their unusually high reflection amplitudes of positive phase (white strips) and lin- ear geometries. Surface cracks are best seen in B- or D-scans and recognized by near-surface hyperbolas. Any unusual feature detected in the radar scans is hereafter referred to as an anomaly. Anomalies are usually reflections of significant amplitude or extent, where reflections from the geometrical boundaries are not expected. For example, in Figure Q.7d the area in the C-scan exhibiting unusually high amplitudes with reverse phase is designated as an anomaly. Ultrasonic Echo Basic Principles of Operation In this test, a single ultrasonic transducer was used to gener- ate and detect ultrasonic waves in the structure. Ultrasonic is based on the measurement of propagation time to localize cracks, voids, and deteriorations, as well as the thickness of a structure. The speed of ultrasonic pulses propagating through the structure is often correlated to material strength and thus a measure of material quality. Ultrasonic echo was employed here to inspect tunnel linings to estimate the thickness of the (a) (b) Figure Q.5. ZFP scanner packed in its custom-made box (a) and on site in Chesapeake Bay Bridge-Tunnel with ultrasonic echo transducer mounted on it (b).

242 (a) (b) Figure Q.6. 1.5-GHz ground-coupled GPR antenna mounted on the scanning system (a) and close-up view (b). anomaly C-Scan crack D-Scan reinforcement C-Scan (a) dowels C-Scan (b) (c) (d) Figure Q.7. Typical radar scans illustrating (a) reinforcement, (b) dowels, (c) surface crack, and (d) anomaly.

243 lining and detect/locate defects and anomalies within the lin- ing. In the absence of ground truth data, the wave-speed of the lining material was either assumed or estimated from sur- face measurements. As such, the thickness of the tunnel lining as well as the depth of the reflectors could only be approxi- mated. Ultrasonic data collection was done automatically using the previously described scanning system. As data collection was conducted point-by-point following a predefined grid, the resulting signals (A-scans) were pro- cessed and presented in real time as evolving B-, C-, and D-scans. Figure Q.8 shows typical A- and B-scans obtained in one of the tunnels. The A-scan shows the intensities of the reflections over time for each point of measurement. The evolution of the A-scans along the profile can be seen in the B-scan. Heterogeneities are recognized by their high reflec- tion amplitudes. Knowing the wave propagation speed made it possible to estimate the depth of the reflector, which could be the tunnel lining backwall or defects. The collected data set could be further processed using the synthetic aperture focus- ing technique (SAFT) algorithm (Schickert et al. 2003) to give a clearer image (higher signal-to-noise ratio) of the internal structure of the test volume; see Figure Q.9). The phase diagram shows the change in phase of ultrasonic waves at the detected interfaces within the material. The color- coded local phase diagram helps distinguish between the reflections from steel objects and from air interfaces. Relative to concrete, steel is of higher and air is of lower impedance. Therefore, the phase of the reflected waves at concrete-steel interfaces and concrete-air interfaces are different. This can be seen in Figure Q.9 where a 180° phase shift (red color) is visible at the location of the rebars, while the backwall reflections exhibit negative phase shifts of 0° to -100°. The advantage of ultrasonic testing is its potential to detect different types of defects such as voids, cracks, honeycombs, B-ScanA-Scan Backwall Figure Q.8. Typical A-scan and B-scan along Chesapeake Bay Bridge-Tunnel lining. B-Scan SAFT B-Scan phase Reinforcement Backwall Figure Q.9. Postprocessed ultrasonic data using SAFT, including both amplitude and phase evaluation.

244 and delaminations directly or indirectly, in real time. Further- more, it can estimate the strength of the material and estimate the structure thickness. Conventional ultrasonic equipment is available and fairly inexpensive. The main limitation is that the transducers must be in contact with the surface of the struc- ture, which slows down the speed of the automated scanning systems. The ultrasonic equipment used by BAM was the A1220. It is a low-cost, multisensor, dry-contact, low-frequency, shear wave transducer developed by Acoustic Control Sys- tems, Ltd. in cooperation with BAM. It includes 24 spring- mounted ultrasonic transducers with a nominal frequency of 50 kHz, out of which 12 serve as transducers and the other 12 as receivers. This construction ensures that a higher amount of ultrasonic energy is transmitted and the reflected and recorded signals can be averaged, thereby minimiz- ing the scatter noise. The images in Figure Q.10 depict the A1220 being used as a handheld device and mounted on the ZFP scanner. For postprocessing of the data (i.e., analy- sis, SAFT, and phase evaluation), two different programs were used: one program was developed at BAM by Rüdiger Feldmann and the other program at the University of Kassel by Dr. Klaus Meyer. Typical Results Ultrasonic echo was able to detect the backwall of tunnel lin- ings directly, reinforcement directly, possible delamination directly and indirectly, surface cracks indirectly, and intru- sions in an otherwise homogeneous volume directly (see Fig- ure Q.11). The indirect detection of reflectors was possible by the “shadow effect,” that is, by recognizing a missing portion of an otherwise consistent element of the tunnel lining, such as backwall or reinforcement. Impact Echo Basic Principles of Operation IE involves introducing a stress pulse into concrete, commonly by application of a mechanical impact on the surface of the structure. A broadband transducer located on the surface close to the impact source (usually at a distance of 2 in. to 4 in.) is used to record vertical deformations of the surface caused by the arrival of incident and reflected waves (or echoes). The response of solid or delaminated plates subjected to IE testing is distinctly different: thickness resonance vibra- tions in case of solid plates, and plate-like flexural vibrations in the presence of shallow, severe delaminations (Shokouhi 2005). Intermediate conditions result in a response super- imposing these two. The time and frequency characteristics of the recorded response can be used to deduce the condition of the struc- ture. Figure Q.12 shows two typical time signals (top) and frequency spectra, one corresponding to a supposedly sound area of a tunnel lining with the backwall as the only reflector with a frequency peak at 3.15 kHz, and the other on a sup- posedly delaminated area, where the spectrum has multiple peaks. The peak in the frequency spectrum of the suppos- edly sound area is the resonance frequency of the tunnel lining depth. The dominant response of a severely delami- nated deck to an impact is characterized by a low-frequency response because of oscillations of the upper delaminated portions of the deck. This response is almost always in the audible frequency range. Because it is significantly lower (a) (b) (c) Figure Q.10. Ultrasonic A1220 from Acoustic Control Systems (a) used for point measurements by hand (b) and used for profile measurements by automated scanning device (c).

245 D-Scan Anomaly D-Scan Reinforcement Backwall Weak backwall echo C-ScanD-Scan Reinforcement Anomaly: delamination? Backwall missing portion of the backwall echo Crack (a) (b) (c) (d) (e) Figure Q.11. Typical ultrasonic echo results: (a) backwall and reinforcement in D-scan, (b) unknown detection of a reflector in a D-scan, (c) direct detection of a delamination- like anomaly in a D-scan, (d) direct detection of delamination-like anomaly in a C-scan, and (e) indirect detection of surface crack considering missing reinforcement in the C-scan. Frequency domain A-Scan time domain time domain Frequency domain A-Scan (a) (b) Figure Q.12. Sound area of tunnel lining with single frequency peak at 3.15 kHz (a) and supposedly delaminated area with several low- frequency peaks (b) from Chesapeake Bay Bridge-Tunnel, Area 3. than the return frequency for the tunnel lining backwall, it produces an apparent reflector depth that is larger than the full thickness (Shokouhi 2005). When using automated scanning devices, the single point- by-point measurements along a profile add up to B- and C-, and D-scans as time/depth/frequency slices (Figure Q.13). Typical Results IE is best known for thickness evaluation and delamination detection in plate-like structures. Depending on the mechanical source used, shallow or deep structures or defects may be inves- tigated. IE can detect the backwall, even at delaminated areas, provided that the delaminations are not severe. Figure Q.12a shows an example of a sound tunnel lining with the backwall resonant single frequency peak at 3.4 kHz. Considering the P-wave velocity of about vP in concrete, ~4,000 m/s, this amounts to a depth of about ~2 ft. Frequency domain D-scans along two selected test lines are shown in Figure Q.13. The D-scan in Figure Q.13a was taken over a sound area, and the D-scan in Figure Q.13b was taken across a supposedly delam- inated area.

246 Measurement Results for eisenhower Memorial Tunnel Description of the Test Area Measurements in the Eisenhower Memorial Tunnel were performed over two days: October 3 and 4, 2011. The unusu- ally high elevation of the tunnel (11,112 ft) created a few challenges. The ZFP scanner is attached to the surface using four vacuum feet. Because of the high tunnel altitude, the compressor could not maintain the pressure necessary to create the vacuum under the feet. Two areas were exam- ined. The first area was regarded as an equipment test. The second area was properly examined with all three NDT methods. The second test area was located within Segment 11, Block 1 of the Eisenhower Memorial Tunnel and was tested from east to west, starting in the lower right corner, then moving up and left. This test area is hereafter referred to as EH2. As shown in Figure Q.14, the 40-in. by 24-in. test area was located 26 in. east of a joint between Segments 10 and 11. The starting scanning point was located at the lower right corner of the scanner field. The scanner moved up and then left, in this case from east to west. The feet of the scanner slid down 0.5 in. during the testing (because of the difficulties in D-Scan D-Scan delamination line x f (a) (b) Figure Q.13. Frequency domain D-scans of test line on sound portion of one tunnel lining with backwall at 3.4 kHz (a) and on supposedly delaminated area of another tunnel lining (b). Figure Q.14. Sketch of test area EH2, located within Segment 11, Block 1.

247 maintaining suction). Figure Q.15a shows the setup on the first area, and Figure Q.15b is a photograph of a page of an information bulletin that shows a cross section of the tunnel with the two fields of testing marked on it. The grid spacing was 1 in. for ultrasonic echo and IE tests and 2 in. for GPR. The position of the test apparatus and the feet of the scanner were marked with chalk. The lengths of the ultrasonic echo, IE, and GPR transducers were parallel to the centerline of the tunnel. The results obtained from each of the three tests performed are discussed below. GPR Results With the GSSI 1.5-GHz GPR antenna, the rebar mesh at a depth between 2 in. and 3 in. was clearly detected. Moreover, an anomaly (reflector of unknown origin) at 16-in. deep was detected. According to the GPR results, the reinforcement bars within the test area along the y-direction were positioned regularly at 10-in. intervals: x = 8 in., x = 18 in., x = 28 in., and x = 38 in. The ones along the x-direction had a 10-in. distancing as well, located at y = 10 in. and y = 20 in. The detected anomaly had an x-dimension of 10 in. extending from x = 18 in. to x = 28 in. and ran along the entire y-dimension of the test area. Figure Q.16 is a three-dimensional (3D) image of the volume with the slices (B-, C-, and D-scans) positioned to reveal the anomaly and the reinforcing ele- ments. Figure Q.17 shows a selection of the B-, C-, and D-scans from EH2, including the detected reflectors: (a) is a D-scan of the reinforcement bars in the y- and x-direction as well as the anomaly, (b) and (c) are B-scans, and (d) and (e) are C-scans showing the reinforcement bars and the anomaly at their respective depths. Ultrasonic Echo Results Using ultrasonic echo, reinforcements at a depth between z = 2 in. and z = 3 in. were detected. An anomaly (i.e., a reflector of an unknown origin) was also detected at an approximate depth of 16 in. The reinforcement bars along the y-direction were 10 in. apart, located at x = 8 in., x = 18 in., x = 28 in., and x = 38 in. The bars along the x-direction were not very clear Figure Q.15. Test area in Eisenhower Memorial Tunnel: (a) test setup of scanner on EH2 and (b) cross section of tunnel showing locations of the two test areas. (a) (b)

248 (because of the positioning and polarization of the probe) and thus could only be vaguely traced at y = 10 in. (Fig- ure Q.18). Figure Q.19 is a 3-D image of the volume with the slices (B-, C-, and D-scans) positioned to reveal the anomaly. The local phase at rebar reflections was, as expected, between 90° and 180° (red color), indicating an impedance higher than the surrounding concrete. The anomaly was 10 in. in width and lay between x = 18 in. and x = 28 in. in the direction of x, as seen in Figure Q.20. It ran completely across the y-dimension of the test area. The local phase was negative, between 0° and -110° (green, yellow), indicating an imped- ance lower than concrete. Impact Echo Results IE could not register either reinforcement or the anomaly detected by GPR and ultrasonic echo. The frequency spectra did not have clear amplitude peaks but was rather a pla teau of many overlapping peaks (Figure Q.21). IE could not yield any reliable information about the backwall of the tunnel lin- ing. Assuming a shear wave velocity of 3,400 m/s, the longi- tudinal wave velocity was approximately 5,889 m/s. Comparison of Results As expected, GPR proved to be the best tool in identifying and locating the reinforcement within the EH2 test volume. Ultra- sonic echo, however, could locate the anomaly of unknown origin more clearly. The negative local phase of the ampli- tudes at the mysterious reflector led to the assumption that the anomaly would have a lesser impedance than the sur- rounding concrete. The fact that GPR registered the anom- aly at a depth of 16 in. led to speculations about it having a higher impedance than the surrounding concrete. Wood Figure Q.16. EH2, GPR: 3-D image of volume positioned to reveal anomaly, with B-scan (a) positioned at x  8 in., C-scan (b) at z  16.8 in., and D-Scan (c) at y  18 in. (b) (c)(a)

249 z = 3 inches Reinforcement z = 16 inches Anomaly Reinforcement y-direction x = 8 inches Anomaly Reinforcement x-direction x = 24 inches Reinforcement y-direction Anomaly y = 4 inches Reinforcement x-direction (a) (b) (d) (e) (c) Figure Q.17. EH2, GPR: Collection of B-, C-, and D-scans from test area displaying main GPR results. and air voids in such a depth could not likely be seen with a 1.5-GHz antenna as clearly as the anomaly seen in the radargrams. One hypothesis is that the anomaly could be one of the steel beams shown in Figure Q.22. This assumption raises the question of how the local phase normally associated with reflections from wood or air could appear. One theory is that the concrete around the metal beams may not be properly bonded to the steel anymore, leaving a thin layer of air between the two mediums. The local phase reflects the phase shift at the concrete-air interface rather than the steel underneath. In general, both the GPR and ultrasonic echo methods were effective in detecting reflectors within the Eisenhower Memorial Tunnel lining. A combination of the two result sets would provide the most detailed and reliable results. Both methods detected the reinforcement and an unknown anom- aly. GPR was more effective in detecting the former and ultra- sonic echo in detecting the latter. The backwall could not be seen with any of the employed techniques here.

250 (a) (c) x z y (b) Figure Q.18. EH2, ultrasonic echo: 3-D image of volume positioned to reveal anomaly, with B-scan (a) positioned at x  19 in., 6-in. width; C-scan (b) at z  14 in., 3-in. width; and D-scan (c) at y  4 in., 2-in. width. y x z = 3 inches Reinforcement y-direction Reinforcement y-direction y z x = 9 inches Figure Q.19. EH2, ultrasonic echo: B- and C-scans, displaying reinforcement. Reflector at 0.7 m could not be identified, as it showed up inconsistently. Test Area 1 Description of the Test Area The first test area was situated in Section 56 of the tunnel and is hereafter referred to as HL1. No referencing system for this tunnel was available. The test area was believed to be within Section 56 because 56/57 was marked with spray paint on the floor, looking south on the right side of the test location (see Figure Q.23a). On the left side, E16 was written. A repaired crack ran across the selected test area (see Figure Q.23b). As shown schematically in Figure Q.24, the 48-in. by 24-in. test area was located between two joints at a distance of 27 in. from the north joint and 56 in. from the south joint. The scanning started at the point closest to the centerline. It then moved away from the centerline and farther south, toward the tunnel entrance. The grid spacing was 1 in. for ultrasonic echo and IE testing and 5 in. for GPR testing. The length (larger dimension) of the ultrasonic echo, IE, and GPR trans- ducer was parallel to the centerline of the tunnel. GPR Results Using the GSSI 1.5-GHz GPR antenna, the reinforcement mesh and other reinforcing elements (possibly dowels) at depths (z) between 1 in. and 6 in. could be detected, as well as an extended anomaly at z = 12 in. Figure Q.25 is a 3-D image of the volume with the slices (B-, C-, and D-scans) positioned to reveal the anomaly and the reinforcing elements. At a Measurement Results for Hanging Lake Tunnel Testing in the Hanging Lake Tunnel took place on October 5 and 6, 2011. As noted in Figure Q.23, three test areas were measured with GPR, ultrasonic echo, and IE.

251 Figure Q.22. Photograph of construction of Eisenhower Memorial Tunnel sometime between 1968 and 1973 (Colorado DOT 2011). Steel beams could be anomaly seen in GPR and ultrasonic echo data. depth of 1 in., rebar-like reflections at x = 24 in. were detected. At this depth, reflections also appeared from a series of shorter elements (dowels) between the two rebar-like reflec- tions. The shorter elements were regularly spaced and ori- ented perpendicular to the rebar-like reflections. The C-scan in Figure Q.26c shows the reinforcement mesh in both direc- tions, with the bars along the y-direction at x = 6 in. and x = 38 in. The D-scans in Figure Q.27 and the B-scan in Figure Q.28 show the third rebar, along the y-direction positioned at x = 22 in., whose reflections could not be distinguished in the C-scans because of the overlap with other reinforcing elements present. The depth of the rebars along the y-direction was between z = 5 in. and z = 6 in. Besides the reinforcing elements, at a depth of z = 5.5 in., an anomaly appeared from z = 12 in., down to z = 16 in., which Figure Q.20. EH2, ultrasonic echo: B-, C-, and D-scans and corresponding local phase diagram, displaying a reflector. z = 16 inchesx = 22 inchesy = 5 inches z = 16.4 inches y z y x x Figure Q.21. EH2, IE: A- and D-scans. No information about possible anomalies, reinforcements, or lining thickness could be drawn. A-Scan x = 25 inches x f

252 Figure Q.24. Sketch of test area, HL1, Segment 56. Figure Q.23. Mark on tunnel vent floor of HL1 used as reference to identify sections (a) and location of repaired crack across test field in relation to scanner aperture (b). (a) (b)

253 led to the rise of amplitudes over an area (Figure Q.26d, Fig- ure Q.27c, Figure Q.28b, and Figure Q.28c). The backwall could not be seen in the GPR radargrams. Ultrasonic Echo Results The reinforcement bars could not clearly be detected in ultra- sonic results. No usable C-scan at the depth of the reinforce- ment. And no horizontal (the same depth) backwall could be identified. However, a deeper reflector plane (relative to the rebar mesh) of variable depth could be detected. Figure Q.29 shows a D-scan taken at y = 5 in. along with the corresponding phase diagram. The reinforcement was seen vaguely at depths between z = 4 in. to z = 6 in. However, an anomalous reflector of mostly negative phase shift appeared at a depth between z = 12 in. and z = 16 in. The B-scans at x = 6 in. and x = 23 in. are shown in Fig- ure Q.30. Figure Q.31, a 3D image of the volume, gives another view on the anomaly. Impact Echo Results The IE spectra contained peaks of frequencies much higher than the expected backwall resonance frequency. Figure Q.32 is a D-scan cut through the short side of the rectangular test area, therefore along the length. Many fre- quency peaks occurred within the frequency spectrum. The first peak of the individual spectra appeared at around 3,700 Hz and 3,400 Hz, corresponding to depths of z = 22 in. and z = 24 in., respectively (assuming a longitudinal wave velocity of 4,000 m/s). A recurring second peak occurred at about 6,700 Hz, corresponding to a shallower reflector, at about (b) (a) (c) Figure Q.25. HL1, GPR: 3-D image of volume positioned to reveal anomaly and reinforcing elements, with B-scan (a) positioned at x  22 in., C-scan (b) at z  12 in., and D-scan (c) at y  1 in.

254 z = 12 in. Because the amplitude spectrum along the profile seemed rather scattered, no reliable conclusions about the nature of the reflector could be drawn. Comparison of Results For this test area (HL1), GPR proved to be the only method to identify the reinforcement mesh and the reinforcing elements. The fine measurement grid and 3-D data collection allowed detection of reinforcing elements overlapping each other in some views. The ultrasonic echo technique, however, was able to detect a deeper anomaly and establish that the anomaly under the test area is located at different depths. The phase diagram provided some information about the possible nature of the anomaly, which appears to have a lower imped- ance relative to its surrounding concrete. IE spectra contained high-frequency energy, but no reliable information could be extracted from either IE time histories or spectra. Comparing the results reveals the need to employ at least two complementary NDT techniques to locate different reflectors within the tunnel lining. GPR is best for locating the metallic reflectors within the penetration range of the antenna. To locate reflectors of different acoustic impedance such as voids and delaminations, the acoustic wave meth- ods should be used. A change of structure seems to occur in the middle of the test area; unusual dowel-like reinforcing elements are present around this location and the depth of the detected anomaly changed abruptly in the ultrasonic results. None of the NDT techniques that were used were able to reliably identify the extent of the tunnel lining. Obviously, the impedance contrast between the tunnel lining and the sur- rounding rock formations was not detectable. The backwall was located outside the penetration range of GPR and pos- sibly the zone of influence of ultrasonic echo. Moreover, the reflection and scattering effects due to the presence of rein- forcement and anomalies weaken the propagating wave and limit its penetration depth. Test Area 2 Description of the Test Area The second test area at Hanging Lake Tunnel (HL2) took place in Segment 57 (segment number was assumed based on the marking on the ground as shown in Figure Q.23a). The 48-in. by 24-in. test area was located 52 in. north of the joint (a) (b) z = 1 in. 0.00 0.50 1.00 -0.60 -0.40 -0.20 -0.00 (c) Reinforcement z = 5.5 in. 0.00 0.50 1.00 -0.60 -0.40 -0.20 -0.00 (d) z = 12 in.Anomaly 0.00 0.50 1.00 -0.60 -0.40 -0.20 -0.00 Dowels Two beams z = 3 in. 0.00 0.50 1.00 -0.60 -0.40 -0.20 -0.00 Figure Q.26. HL1, GPR: C-scans with (a) rebar-like reflection, (b) reflections from two rebars appearing close to each other with dowel-like elements in between along with shallower part of reinforcement bar in x-direction seen at y  12 in., (c) reinforcement mesh in both directions, and (d) anomaly.

255 Reinforcement, y-direction y = 22 in. Reinforcement, x-direction Area of beams and dowels y = 12 in. Anomaly y = 6 in. (a) (b) (c) Figure Q.27. HL1, GPR: D-scans. Area of beams and dowels (a) as well as reinforcement in x- and y-direction can be seen as a cut through. Reinforcement in x-direction starts at z  3 in. and leads to z  5 in. Third reinforcement bar along y-direction at x  22 in. is located beneath other reinforcing elements at x  24 in. (b), which makes it difficult to distinguish them in C-scan. Image (c) shows anomaly starting from z  12 in. and going to z  16 in. Image has higher gain than other images to clarify the anomaly.

256 Reinforcement, y-direction Anomaly y = 5 in. z = 16 in. z = 12in. x z Figure Q.29. HL1, ultrasonic echo: D-scan at y  5 in. Curved anomalous reflector of mostly negative phase detected between z  12 in. and z  16 in. x = 6 in. x = 22 in. x = 38 in. x = 34 in. Reinforcement along x-direction Reinforcement along y-direction Dowels Reinforcement along x-direction Anomaly Anomaly (a) (b) (c) (d) Figure Q.28. HL1, GPR: B-scans. First reinforcement bar along y-direction at x  6 in. (a). At x  34 in., bar along x-direction positioned at y  12 in. (b) as well as anomaly at z  16 in. is seen. Second bar along y-direction at x  22 in. under other reinforcing elements and anomaly at z  12 in. (c), and third bar along y-direction at x  38 in. (d) are also shown.

257 dividing segments 56 and 57. Its upper edge (toward the tun- nel crest) was about 100 in. from the centerline. Figure Q.33 shows a sketch of HL2. An unrepaired crack ran across the test area, as shown in Figure Q.34a. Figures Q.34a and Q.34b, show the ZFP scanner. The scanning started at the point clos- est to the centerline, first moving down and away from the centerline, then south toward the tunnel entrance. The grid spacing for both the ultrasonic echo and IE tests was 1 in. and for GPR was 2 in. The ultrasonic echo, IE, and GPR transduc- ers were mounted such that their length (large dimension) was parallel to the centerline of the tunnel. One profile line of the GPR data was missing (at y = 20 cm, or 8 in.), which caused a discontinuity in the B-scans and C-scans. Conse- quently, no D-scan was available for y = 8 in. GPR Results Figure Q.35 is a 3-D image of the volume intended to give an overall view of the reinforcement. Anomaly Anomaly Reinforcement/beam x = 6 in. x = 23 in. z = 16 in. z = 5 in. z = 12 in. y z (a) (b) Figure Q.30. HL1, ultrasonic echo: B-scans to evaluate extent of anomaly. B-scan crossing through deeper reflector (a) and B-scan crossing through shallower anomaly (b). (b) (a) (c) Figure Q.31. HL1, ultrasonic echo: 3-D image of volume positioned to reveal anomaly, with B-scan (a) positioned at x  6 in.; C-scan (b) at z  21 in., 4-in. width; and D-scan (c) at y  6 in. Figure Q.32. HL1, IE: D-scan (a) and A-scan (b). A-Scan x = 14 in. x f (a) (b)

258 Figure Q.33. Sketch of test area HL2, Segment 57. (a) (b) (c) Figure Q.34. HL2: (a) crack running across test area, (b) NDT scanner mounted on tunnel ceiling, and (c) scanning of test area.

259 In the C-scan at a depth of z = 5 in., a reinforcement bar in the general x-direction could be seen running across the test area roughly from y = 16 in. on the south side of the area to about y = 12 in. on the north side (Figure Q.36b). The D-scan in Figure Q.36c shows that this reinforcement bar ran above the bars perpendicular to it. Figure Q.36a is the B-scan at x = 2 in., where the bar at y = 16 in. was marked. The C-scan at z = 8 in. showed another, albeit weak, reinforcement bar along the x-axis, at about y = 24 in. (see Figure Q.37b). A D-scan through the bar at y = 24 in. (Figure Q.37c) and a B-scan at x = 2 in. (Figure Q.37a), with the weak reflection from the bar marked, are shown as well. The reinforcement bars in the y-direction at a depth of z = 6 in. are shown in Figure Q.38b. Figure Q.38c shows the D-scan at y = 24 in. with the reinforcements marked at a depth of 8 in. The B-scan (Figure Q.38a) through the bar at x = 9 in. revealed that the bars did not run parallel to the surface but were bent from z = 6 in. down to z = 8 in. This explained the weak reflections over half of the C-scans at z = 6 in. Figure Q.39 shows images resulting from the crack on the surface, which manifested itself by a changed impedance because of the intrusion of moisture into the very first layers of the lining. Figure Q.39b is a slice in the depth, this time at z = 3 in. The circled area shows the reflection caused by moisture as a result of the crack on the surface. It ran in that depth until y = 12 in. (y = 30 cm). The upper part until y = 6 in. (y = 15 cm) does not seem to fit onto the lower part. The reason for this is a missing profile line caused by a failure during the measure- ments, which cannot be reconstructed properly by the pro- gram. Figure Q.39c shows the same reflector at a depth of z = 3 in., which does not show up at the other D-scans above, meaning it was local. The B-scan at x = 19 in. (Figure Q.39a) shows the extent of the reflector better: it was down to 3.4-in. deep and nearly 16 in. into the field of measurement. (b) (a) (c) Figure Q.35. HL2, GPR: 3-D image of volume positioned to reveal reinforcing elements, with B-scan (a) positioned at x  44 in.; C-scan (b) at z  6.5 in., 3-in. width; and D-scan (c) at y  0 in.

260 z = 5 in. y = 16 in. x = 2 in. (a) (b) (c) Figure Q.36. HL2, GPR: B-, C-, and D-scans showing first reinforcement bar along x-direction at depth of z  5 in. z = 8 in. y = 24 in. x = 2 in. (a) (b) (c) Figure Q.37. HL2, GPR: B-, C-, D-scans showing second reinforcement bar in x-direction at a depth of z  8 in.

261 y = 24 in. z = 6 in.x = 9 in. (a) (b) (c) Figure Q.38. HL2, GPR: B-, C-, D-scans showing reinforcement bars in y-direction at depths between 6 in. and 8 in. z = 3 in. y = 6 in. x = 18 in. (a) (b) (c) Figure Q.39. HL2, GPR: B-, C-, D-scans showing reflections caused by presence of a surface crack.

262 Ultrasonic Echo The reinforcement mesh could be clearly seen in the ultrasonic echo C-scan at a depth of z = 6 in. (Figure Q.40). The corre- sponding phase diagram is also included in Figure Q.40b. As expected, the local phase of the reinforcement bars’ reflections appeared mostly positive, which represented impedance higher than the surrounding concrete, that is, steel. The B-scan (a) and D-scans (b) in Figure Q.41 show the steel bars in both directions. The reinforcement bar along the x-axis at an approximate depth of z = 8 in. (Figure Q.41c) would not be identified without previous knowledge of its existence through GPR data. The reinforcement bars in the y-direction were, however, easily detectable. The hole in the reinforcing bar reflections in Figure Q.40 and Figure Q.41b are due to the pres- ence of moisture that intruded through the surface crack. It caused the US signal energy to be absorbed. Figure Q.42 is a 3-D image of the volume intended to give an overall view of the reinforcement. Impact Echo Some of IE spectra showed a dominant frequency peak at about 3,600 Hz, which equals a depth of z ~ 22 in. However, in the B- and D-scans, no clear backwall could be seen. Fig- ure Q.43 shows a typical example of the obtained IE D-scan (Figure Q.43a) and A-scan (Figure Q.43b). The IE data for this test area yielded no reliable information about either the thickness of the lining or the presence of possible anomalies. Comparison of Results Reinforcement could be detected using both GPR and ultra- sonic echo, although GPR exhibited a clear advantage in detect- ing deep steel bars, which could not be reliably identified in ultrasonic echo results. The surface crack was seen in the GPR data as a near-surface reflector, maybe because of moisture penetrating the tunnel lining through the crack, resulting in a change of the dielectric constant. The signature of this crack in the ultrasonic echo data was a hole in the reflections from the reinforcement bars. Neither GPR nor ultrasonic echo could give any indication of the thickness of the lining. IE spectra contained a repeated frequency peak at a frequency resonating at a depth of about z ~ 22 in., which could possibly have been the backwall. GPR proved to be the most effective NDT method to detect the reinforcement as well as the effects of a surface crack. Even the extent of the affected area could be detected, as the anomaly was directly influencing the data. The crack could be indirectly detected in the ultrasonic echo results because the reflections from the reinforcements were shadowed by it. It is Reinforcement bar x-direction Reinforcement bar y-direction Shadowed area due to crack Reinforcement bar x-direction x = 44 in. y = 12 in. y = 24 in. x y z z (a) (b) (c) Figure Q.41. HL2, ultrasonic echo: B-scan (a) and D-scans (b) and (c) showing reinforcement in the x-direction (a) and (b) and in the y-direction (c). Reinforcement bar along x-axis at z  8 in. is very weak (c). Shadowed area due to crack z = 6 in. y (a) (b) x Figure Q.40. HL2, ultrasonic echo: C-scan at z  6-in. amplitude (a) and corresponding phase diagram (b). Missing part of reinforcement bar is due to surface crack within test area.

263 a curious finding that the thickness of the lining could not be detected by either method, as the tunnel is relatively recent. IE, however, could provide some hints about the thickness of the tunnel. Ground truth information on the lining thickness would help verify the accuracy of the IE results. Test Area 3 Description of the Test Area The third test area at Hanging Lake Tunnel (HL3) was also near Section 57/58 of the tunnel, which was marked as such with spray on the floor (Figure Q.44a). The test area included a small crack (Figure Q.44b). A transverse joint crossed through the test field, and the area was relatively close to the centerline of the tunnel. The field had an area of 48 in. by 24 in. The distance of the field to permanent features of the tunnel was not measured; therefore, the sketch in Fig- ure Q.45 has no offsets marked. The grid spacing for ultra- sonic echo and IE was 1 in., and for GPR was 2 in. Figure Q.44c shows an image of the ZFP scanner on the test area. During the testing, the longer side (length) of the GPR, ultrasonic echo, and IE transducers was set parallel to the centerline of the tunnel. GPR Results Figure Q.46 shows the reinforcement in the x-direction in B-, C-, and D-scans. One bar is at the edge of the test area. The steel bars ran from around y = 8 in. to y = 6.5 in. at a depth of z = 5 in., and from around y = 24 in. to y = 22.5 in. at a depth of z = 4 in. The transverse joint running across the test field was encircled on the radarscans. The reinforcement bars in the y-direction were positioned at x = 10 in., x = 24 in., and x = 42 in., at a depth of z = 6 in. (Figure Q.47). The scans from the joint crossing the test field at x = 6 in. in z = 2 in. are shown in Figure Q.48. Dowel-like steel elements traversing the joint at z = 3 in. can be seen in the B- and D-scans. The line scans at y = 0.45 cm (18 in.), y = 0.4 cm (16 in.), and y = 0.35 cm (14 in.) were missing. Figure Q.49 is a 3-D image of the volume intended to give an overall view of the reinforcing elements. Ultrasonic Echo Results The reinforcement running along both x- and y-directions at about z = 6 in. could be seen in the ultrasonic echo results, although the rebars along the x-axis at y = 8 in. and y = 24 in. did not appear as clearly as the ones along the y-axis at x = 9 in., x = 24 in., and x = 40 in. (Figure Q.50). The joint could not be identified in the ultrasonic echo results. (a) (c) (b) Figure Q.42. HL2, ultrasonic echo: 3-D image of volume positioned to reveal reinforcing elements, with B-scan (a) positioned at x  10.5 in., C-scan (b) at z  8 in., and D-scan (c) at y  12 in. Missing reinforcement due to crack is also seen in D-scan. A-Scan x = 17 in. x f (a) (b) Figure Q.43. HL2, IE: D-scan (a) and A-scan (b). First frequency peak at 3,600 Hz could represent backwall in depth of z ~ 22 in.

264 (a) (c) (b) Figure Q.44. HL3: (a) section marking on the floor, (b) crack across the field, and (c) ZFP scanner.

265 Figure Q.45. Sketch of test area HL3, segment 58, Hanging Lake Tunnel. z = 4.5 in. y = 8 in. x = 3 in. Reinforcement x-direction Joint y-direction y = 24 in. Reinforcement x-direction (a) (b) (c) Figure Q.46. HL3, GPR: B-, C-, and D-scans of reinforcement in x-direction running along at a depth of z  5 in. and z  4 in. B-scan taken at x  3 in. (a). C-scan at z  4.5 in. with a width of 1.5 in. (b). D-scans taken from points y  8 in. and y  24 in. (c). Circled reflector is the joint.

266 z = 6 in. y = 8 in. x = 24 in. Joint y-direction Reinforcement y-direction (a) (b) (c) Figure Q.47. HL3, GPR: B-, C-, and D-scans of reinforcement bars running along y-direction at a depth of z  6 in. B-scan taken at x  24 in. (a). C-scan at z  6 in. (b). D-scan at y  8 in. (c). Circled reflector is the joint. x = 6 in. z = 2 in. y = 24 in. Joint Dowels Reinforcement x-direction (a) (b) (c) Figure Q.48. HL3, GPR: B-, C-, D-scans corresponding to location of joint at x  6 in.

267 (b) (c) (a) Figure Q.49. HL3, GPR: 3-D image of volume positioned to reveal reinforcing elements, with B-scan (a) positioned at x  8 in.; C-scan (b) at z  7 in., 1 in. wide; and D-scan (c) at y  16 in. z = 6 in. x = 3 in. y = 16 in. (a) (b) (c) y x x z y z Figure Q.50. HL3, ultrasonic echo: B-, C-, D-scans showing reinforcement bars running along x- and y-direction at depth of z  6 in.

268 Impact Echo Results No features were resolved on the basis of the IE results, as the frequency peaks were generally too broad to point to any par- ticular resonating feature (Figure Q.51). Comparison of Results Reinforcement could be detected with both GPR and ultra- sonic echo, but they were much clearer with GPR. The joint seen on the surface of the field could only be detected with GPR, as well as two beams. None of the NDT methods resolved the end of the tunnel lining. Measurement Results for Chesapeake Bay Bridge-Tunnel Measurements in the Chesapeake Channel Tunnel (a part of the Chesapeake Bay Bridge-Tunnel system) were taken over two days: October 11 and 12, 2011. Four test fields were tested, three of which were located in the tunnel’s exhaust air duct (shown in Figure Q.52). The fourth test area was on the tiles of the tunnel wall itself. Test Area 1 Description of the Test Area The first test field (CPB1) was located at Station 474+27. A sketch of the field showing its positioning in the tunnel is shown in Figure Q.53. The 48-in. by 24-in. test area started immediately above the pipe. A joint ran across the field, paral- lel to its shorter side, at one-fourth its length. The center of the ZFP scanner’s feet were at 14 in. to the left of the joint marked Station 474+27, facing the tunnel wall and about 29 in. down from the joint running along the tunnel. When looking north, this joint was about 7 in. to the left of the vertical stands (cen- terline). The tunnel’s entrance is at Station 470 south end. The Figure Q.52. Cross section of exhaust air duct of Chesapeake Channel Tunnel. (a) (b) A-Scan x = 22 in. x f Figure Q.51. HL3, IE: Typical D-scan (a) and A-scan (b). No clear frequency peaks, which could have represented a possible backwall or other reflectors, could be identified.

269 backwall has a nominal thickness of 24 in., according to the tunnel’s blueprints, and has a so-called steel skin. The grid spacing for ultrasonic echo and IE testing was set to 1 in., and for GPR, 2 in. During the testing, the GPR, ultra- sonic echo, and IE transducers were positioned such that their lengths were parallel to the centerline of the tunnel. GPR Results Figure Q.54 gives an overview of the reinforcement within the volume. The reinforcements in the y-direction at a depth of z = 1.5 in. and in the x-direction at a depth of z = 3 in. are depicted in the C-scans of Figure Q.55. The transverse bars along the y-axis were detected at x = 0 in. (i.e., under one of the shorter sides of the test area), x = 12 in., x = 24 in., and x = 42 in. The longi- tudinal reinforcement was detected at y = 16 in. Figure Q.56 includes two D-scans, one at y = 22 in. and another one at y = 16 in. that cuts through the longitudinal rebar appearing in Figure Q.55b. A selection of B-scans cut- ting through the transverse rebars at x = 24 in. and x = 42 in. is shown in Figure Q.57. Ultrasonic Echo Results Transverse steel bars along the y-direction could be clearly detected and identified using the ultrasonic echo technique as well (see the C-scan of Figure Q.58 taken at z = 2 in.). The reflections from the longitudinal bar in the x-direction in the C-scan were vaguely seen, although they could be more clearly seen in the later B-scans. This repeating observation resulted from the orientation of the ultrasonic echo trans- ducer, which made it more sensitive to transverse reinforce- ment (the polarization effect). When examining the C-scans (Figure Q.59) at different depths, the steel skin of the tunnel at about z = 24 in. was detected. The shear wave velocity was adjusted to about 2,710 m/s such that the backwall reflections occurred at the known depth of 24 in. A shallower anomaly was detected at z = 15 in., spreading from x = 20 in. to x = 40 in. and y = 6 in. to y = 14 in., detected both directly and indirectly through its shadowing of the skin reflections. The correspond- ing phase diagrams exhibited a positive phase for the reinforce- ment bars and steel skin and a mixture of positive and negative angles (inconclusive) for the anomaly. Looking at the D-scan at y = 12 in. (Figure Q.60), the rebars and the backwall at z = 24 in. are clearly seen. About one-half of the backwall (with x > 24 in.) exhibited weakened reflec- tions or even missing reflections. The phase evaluation offered more information about why the backwall reflections appeared weaker: at a depth of around z = 16 in. was a con- fined anomaly that shadowed the backwall reflection between x = 22 in. and x = 30 in. This anomaly was the same as seen in the C-scan in Figure Q.59. Figure Q.53. Sketch of test area CPB1, Station 474-27, Chesapeake Channel Tunnel.

270 (b) (a)(c) Figure Q.54. CPB1, GPR: 3-D image of volume positioned to reveal reinforcing elements, with B-scan (a) positioned at x  38 in.; C-scan (b) at z  4 in., 1 in. wide; and D-scan (c) at y  6 in. Reinforcement bars y-direction Reinforcement bar x-direction z = 1.5 in. z = 3 in. (a) (b) Figure Q.55. CPB1, GPR: C-scans depicting reinforcement bars (a) along y-directions at z  1.5 in. and (b) in the x-direction at a depth of about z  3 in.

271 y = 22 in. y = 16 in. Reinforcement bars y-direction (a) (b) Reinforcement bar x-direction Figure Q.56. CPB1, GPR: D-scans showing reinforcement bars (a) along y-direction at location y  22 in. and (b) x-direction at location y  16 in. x = 24 in. x = 42 in. (a) (b) Reinforcement bar y-direction Figure Q.57. CPB1, GPR: B-scans of transverse rebar (along y-direction) taken at (a) x  24 in. and (b) x  42 in. z = 2 in. x y Reinforcement bars y-direction Figure Q.58. CPB1, ultrasonic echo: C-scan showing transverse reinforcement in y-direction at z  2 in.

272 The B-scans in Figure Q.61 were chosen to represent areas of the test volume with and without anomaly. The left image in Figure Q.61 is a B-scan at x = 3 in., with no anomaly. The backwall and reinforcement bar in x-direction (at y = 16 in.) can be clearly seen. The other B-scan was obtained by cutting through the anomaly at x = 28 in. The backwall is missing between y = 8 in. and y = 12 in. and some reflections from about z = 13 in. downwards can be seen. The corresponding phase diagrams are not conclusive. Figure Q.62 is a 3-D image of the volume, showing the backwall and anomaly. Impact Echo Results The IE spectra contained a clear thickness resonance fre- quency peak at a frequency between 3,600 Hz and 3,400 Hz, representing a thickness of about z = 22 in. to z = 24 in., assum- ing a longitudinal wave velocity of 4,000 m/s. Figure Q.63 shows the selected B-scan (Figure Q.63a) and D-scan (Figure Q.63b) of the test volume along with a representative A-scan and fre- quency spectrum. The amplitudes of the backwall echo in the D-scan between x = 27 in. and x = 46 in. were weaker than in the area x < 27 in., indicating the presence of an inhomogene- ity between the sensor and the backwall absorbing the wave energy. The same was true for the B-scan between y = 0 in. and y = 8 in. Comparison of Results While GPR proved to be the most reliable NDT method for detecting and identifying reinforcement bars, it could not detect a 15-in.-deep localized anomaly. The ultrasonic echo z = 15 in. y x z = 24 in. (b) (a) Figure Q.59. CPB1, ultrasonic echo: C-scans showing reflection from anomaly (a) and tunnel skin (b). y = 12 in. x z Backwall and weaker backwall reflection Figure Q.60. CPB1, ultrasonic echo: D-scan showing reinforcement and tunnel skin. Weakened reflection from backwall is marked.

273 x = 3 in. x = 28 in. Backwall Reinforcement bar x - direction Backwall Anomaly (a) (b) y z z y Figure Q.61. CPB1, ultrasonic echo: B-scan showing longitudinal reinforcement bar (along x-direction) and backwall, where no anomaly was detected (a), and where an anomaly was detected (b). (b) (a) (c) Figure Q.62. CPB1, ultrasonic echo: 3-D image of volume positioned to reveal backwall and anomaly, with B-scan (a) positioned at x  29 in., C-scan (b) at z  24 in., and D-scan (c) at y  7 in. technique, however, was not as clear in detecting the steel bars (due to the polarization effects) but indicated the presence of an anomaly, directly and indirectly (directly by evaluating the reflections from the anomaly and indirectly based on the weakened and even missing backwall echo). Both ultrasonic echo and IE could yield the thickness of the tunnel lining. The phase diagrams allowed the ultrasonic echo results to even indicate that the impedance of the tun- nel’s skin was higher than that of the lining and, therefore, of steel. This could be verified by the tunnel’s blueprint indicat- ing the presence of a steel skin. Considering the obtained data from all three employed NDT methods together, a clearer picture of the geometry and condition of the tunnel emerged. Test Area 2 Description of the Test Area The test area CPB2 was located on the west side of the tunnel, facing north. It was located south of the joint marked Station 481+76, and the entire area was on a single block. The 48-in. by 24-in. test area was oriented such that its length was parallel to the centerline of the tunnel. The near and far shorter sides of the scanning aperture were 9 in. and 49 in. south of Station 481+76. The longer side was 47 in. away from the tunnel’s centerline, where the vertical stands were. Figure Q.64 provides a rough sketch of the test area and its positioning within the tunnel. The grid spacing for ultrasonic echo and IE testing was set to 1 in., and for GPR, it was 2 in. The GPR, ultrasonic echo, and IE transducers were oriented such that their lengths were par- allel to the centerline of the tunnel. Note that the GPR mea- surements for this test area were shifted 1.5 in. in the x-direction compared with those from IE and ultrasonic echo. GPR Results Figure Q.65 is a 3-D image of the volume with the slices (B-, C-, and D-scans) positioned to reveal the reinforcing elements. Transverse reinforcement in the y-direction could be detected at a depth of z = 2.5 in. with the bars positioned at x = 0 in., x = 12 in., x = 24 in., and x = 36 in. Two longitudinal steel rebars in the x-direction appeared 4-in. deep at y = 8 in. and y = 24 in. The one at y = 24 in. ran out of the test field and was only partly visible. Figure Q.66 provides C-scans showing the reinforcement mesh at two different depths. D-scans taken at y = 8 in. and y = 24 in. showed the rebars in the y-direction as well as one of the rebars in the x-direction (see Figure Q.67). One B-scan at x = 12 in. cutting through one of the transverse rebars and a second one at x = 16 in. cutting between the transverse rebars, showing the longitudinal rebars, are shown in Figure Q.68.

274 (a) (b) A-Scan y = 5 in. A-Scan x = 38 in. x f f y A-Scan y = 12 in. A-Scan x = 4 in. Figure Q.63. CPB1, IE: (a) B-scan and (b) D-scan of profile lines along y- and x-axes of field coordinate system. A-scans (right ): Frequency peaks represent tunnel lining thickness. Areas with weaker amplitudes indicate wave energy absorbing inhomogeneity. Figure Q.64. Sketch of test area CPB2 south of Station 48176, Chesapeake Channel Tunnel.

275 (a) (b) (c) Figure Q.65. CPB2, GPR: 3D image of volume positioned to reveal reinforcing elements, with B-scan (a) positioned at x  36 in.; C-scan (b) at z  4 in., 4 in. wide; and D-scan (c) at y  8 in. z = 2.5 in. z = 4 in. Reinforcement y-direction Reinforcement x-direction (b)(a) Figure Q.66. CPB2, GPR: C-scans at z  2.5 in. and z  4 in. showing rebars in y-direction (a) and x-direction (b).

276 Ultrasonic Echo Results Figure Q.69 shows a 3-D image of the volume focusing on the backwall and an anomaly shadowing it. The reinforcement mesh was not clear in the ultrasonic echo C-scans. The transverse bars along the y-axis were most clearly seen in the C-scan at z = 2 in. (Figure Q.70a), and those along the x-axis at z = 4 in. (Figure Q.70b). Besides the reinforcements, an anomaly at z = 20 in. (Fig- ure Q.71a) and the backwall at z = 28 in. (Figure Q.71b) could also be detected in the C-scans. The anomaly had a phase shift between 45° and -45° (of lower acoustic impedance than concrete). The backwall exhibited a phase shift between -90° and -180° (of higher acoustic impedance than concrete). y = 8 in. y = 24 in. (a) (b) Reinforcement y-direction Reinforcement x-direction Figure Q.67. CPB2, GPR: D-scans through test volume at positions y  8 in. and y  24 in. showing reinforcement in y-direction (a) and in x-direction (b). x = 12 in. x = 16 in. Reinforcement y-direction Reinforcement x-direction (a) (b) Figure Q.68. CPB2, GPR: B-scans taken at x  12 in. and x  16 in. showing rebars in y-direction (a) and in x-direction (b). To analyze the ultrasonic echo data, a transversal wave veloc- ity of 2,710 m/s was assumed (taken from the measurements at CPB1). By examining D-scans (Figure Q.72) and B-scans (Fig- ure Q.73), the extent of the anomaly could be approximated as running along the entire width of test field CPB2 (i.e., y-direction) but confined between x = 8 in. and x = 28 in. in the x-direction. Impact Echo Results The IE spectra contained two distinct frequency peaks at about 2,900 Hz and 4,100 Hz (see spectral A-scans of

277 Figure Q.74) corresponding to the reflector depths of z = 27 in. (Figure Q.74a) and z = 20 in. (Figure Q.74b), respectively. The D-scan in Figure Q.75 shows a clear shift in the fre- quency peak from 3,000 Hz up to 3,700 Hz, corresponding to the approximate depths of z = 27 in. (lining thickness) and z = 21 in. (anomaly). Comparison of Results Steel reinforcement was best located with GPR. Two reflectors at two different depths were detected by both ultrasonic echo and IE. Phase analysis of the ultrasonic echo results showed dif- ferent phase shifts at the reflectors: the phase shift at the shal- lower reflector (at 20 in.) indicated an impedance lower than concrete, while the deeper reflector (at 28 in.) had an acoustic impedance higher than concrete. This higher impedance is typical of concrete-metal interfaces and leads to the assumption of it being the echo of the metal skin surrounding the tunnel, although the estimated depth of 28 in. does not match the (b) (a) (c) Figure Q.69. CPB2, ultrasonic echo: 3-D image of volume positioned to reveal backwall and anomaly, with B-scan (a) positioned at x  3 in.; C-scan (b) at z  25 in., 5 in. wide; and D-scan (c) at y  4 in. (a) z = 4 in.z = 2 in. (b) x y Figure Q.70. CPB2, ultrasonic echo: C-scans at z  2 in. and z  4 in. showing the reinforcement along (a) y- and (b) x-directions. (a) (b) z = 20 in. z = 28 in. x y Figure Q.71. CPB2, ultrasonic echo: C-scans at different depths: (a) depth of anomaly at 20 in. and (b) depth of tunnel lining at 28 in.; image of amplitudes ( left ), image of corresponding phase (right ).

278 y = 14 in. Reinforcement y-direction A B x z Figure Q.72. CPB2, ultrasonic echo: D-scan at y  14 in. and corresponding phase diagram. x = 2 in. x = 9 in. x = 36 in. anomaly backwallbackwall Reinforcement y-direction Reinforcement x-direction z y Figure Q.73. CPB2, ultrasonic echo: B-scans through anomaly and backwall taken at x  2 in., x  9 in., and x  36 in.

279 nominal thickness of 24 in. This could be a result of errors in the assumed concrete shear wave velocity used in esti- mating the reflector depths to analyze the US data of CPB2, as the shear wave velocity was assumed to be 2,710 m/s (from CPB1). The use of the GPR and at least one acoustic NDT method was necessary to analyze CPB2. GPR could reliably detect and identify the reinforcement bars. Both acoustic methods detected the echoes from an anomaly and the backwall. The accurate estimation of the reflector depth was possible only when the wave velocities at test locations were known. Neither ultrasonic echo nor IE could provide the wave velocity of the test medium without having ground truth information. To measure the velocity profile in situ, other methods such as high frequency spectral analysis of surface waves or multispectral analysis of surface waves could be employed. Test Area 3 Description of the Test Area As seen in the sketch in Figure Q.76, a joint near Station 486+67 ran almost through the middle of the test area, hereafter referred to as CPB3. The area was on the east side of the tunnel, opposite the two other test fields (CPB1 and CPB2). The test area was 48 in. long and 24 in. wide, (a) A-Scan y = 13 in. A-Scan y = 17 in. (b) y f y f Figure Q.74. CPB2, IE: Spectral A-scans taken at two different locations: (a) through anomaly and (b) through backwall. A-Scan x = 1 in. A-Scan x = 10 in. f x Figure Q.75. CPB2, IE: Spectral D-scan (along x-axis) along with two selected spectral A-scans representing different depths.

280 extending 22 in. south and 26 in. north of the joint. The test field started at an offset of 24 in. from the centerline of the tunnel. The grid spacing for ultrasonic echo and IE testing was 1 in. and for GPR testing was 2 in. The length of the GPR, ultra- sonic echo, and IE transducers was parallel to the centerline of the tunnel during the scanning. GPR Results The reinforcement mesh could be easily seen in GPR C-scans. The transverse rebars in the y-direction were spaced 12 in. from each other and were positioned at x = 2 in., x = 14 in., x = 26 in., and x = 38 in. at a depth of z = 1.6 in., as shown in Figure Q.77. The longitudinal ones in the x-direction were at y = 0 in., y = 18 in., and z = 3 in.-deep (see Figure Q.77). The D-scans taken at y = 22 in. and y = 18 in. (Figure Q.78) clearly showed that the rebars in the y-direction ran above those in the x-direction. The B-scan taken at x = 26 in. (Figure Q.79) showed the cross section of one of the transverse rebars (along the y-direction). The one taken at x = 16 in. revealed only longi- tudinal rebars (along the x-direction). Figure Q.80 is a 3-D image of the volume as a summary of the reinforcing elements. Figure Q.76. Sketch of CPB3, Station 48667 Chesapeake Channel Tunnel. Reinforcement y-direction Reinforcement x-direction z = 1.6 in. z = 3 in. (a) (b) Figure Q.77. CPB2, GPR: C-scans taken at (a) z  1.6 in. showing reinforcement bars in y-direction and (b) z  3 in. showing rebars in x-direction.

281 Ultrasonic Echo Results Three of the four transverse reinforcement bars (in the y-direction) could be detected in the ultrasonic echo C-scan at z = 2-in. deep, at x = 2 in., x = 14 in., and x = 38 in. (Fig ure Q.81). An anomaly was present in the middle of the field starting at a depth of z = 2 in. Examining deeper C-scans confirmed the existence of an anomalous reflector (Figure Q.82). At a depth of about z = 25 in. (Figure Q.83), the backwall with a positive phase (indicating an impedance higher than the surrounding con- crete) could be detected. Between x = 16 in. and x = 38 in., the backwall echo was missing because of the shadowing effect of the earlier-described anomaly reflector. To analyze the ultra- sonic echo data of CPB3, the shear wave velocity 2,710 m/s of CPB1 was assumed. A D-scan taken at y = 16 in. is shown in Figure Q.84. This view reveals multiple reflections from the anomaly at z = 6 in., z = 10 in., z = 15 in., and z = 20 in. with changing phase. The backwall echo was missing because the anomaly shadows the deeper reflectors. The multiple reflections with their chang- ing phase gave indications of only one anomaly. However, this could not be verified. A linear reflector of unknown origin was observed at x = 26 in. at the same depth as the backwall (marked with a question mark in both Figure Q.84 and Figure Q.85). Note that at x = 26 in., the joint ran across the test field. As the reflector was not seen in the D-scans produced by the raw data, this was likely an artificial feature produced by the SAFT algorithm. y = 22 in. y = 18 in. Reinforcement x-direction (a) (b) Reinforcement y-direction Figure Q.78. CPB3, GPR: D-scans taken at y  22 in. (a) and y  18 in. (b) reveal reinforcement bars in both directions. x = 26 in. x = 16 in. (a) (b) Reinforcement x-direction Reinforcement y-direction Figure Q.79. CPB3, GPR: B-scans at x  26 in. (a) and x  16 in. (b) reveal reinforcement rebars in both directions.

282 (a) (b) (c) Figure Q.80. CPB3, GPR: 3-D image of volume positioned to reveal reinforcement, with B-scan (a) positioned at x  45 in.; C-scan (b) at z  4.5 in., 4-in. width; and D-scan (c) at y  0 in. Anomaly z = 2 in. x y Figure Q.81. CPB3, ultrasonic echo: C-scan at z  2 in., showing reinforcement bars in y-direction as well as anomaly between x  16 in. and x  38 in. (marked). Corresponding phase diagram (right). z = 4 in. Anomaly x y Figure Q.82. CPB3, ultrasonic echo: C-scan obtained at depth of 4 in. Reflector is confined between x  14 in. and x  36 in.

283 z = 25 in. Missing backwall echo due to the shadowing effect of the shallower anomaly x y Figure Q.83. CPB3, ultrasonic echo: C-scan at z  25 in. reveals backwall echo. Between x  16 in. and x  36 in., echo is missing due to presence of shallower anomaly. y = 16 in. Backwall Missing backwall echo Anomaly Anomaly Reinforcement y-direction x z Figure Q.84. CPB3, ultrasonic echo: D-scan at y  16 in. Backwall can be seen where anomaly does not shadow it. Reinforcement in y-direction is seen at z  3 in. Reflector of unknown origin is positioned at x  26 in. at backwall’s depth. The B-scans did not provide additional information on the anomaly. As seen in Figure Q.85, the backwall of the tunnel lining could be clearly seen at z = 25 in. in the B-scan taken at x = 1.5 in. The transverse rebar in the y-direction was also clearly seen. In contrast, the B-scan at x = 20 in. contained no backwall echoes, as expected from the D-scan, but no further information about the anomaly itself. Figure Q.86 shows a 3-D image of the volume focusing on the backwall and its shadowed area, as well as the shallow anomaly. Impact Echo Results As seen in Figure Q.87, IE showed tunnel lining thickness resonance frequency except between x = 13 in. and x = 30 in., where the echo was disturbed. The typical spectral and tem- poral A-scans from the sound (x = 5.5 in.) and disturbed regions (x = 20 in.) are compared in the figure. While the sound spectrum contained one clearly dominant frequency, the disturbed spectrum contained multiple peaks, mostly of frequencies lower than that of the thickness resonance fre- quency. The thickness resonance frequency appeared at about 3,200 Hz, corresponding to a depth z = 25 in. Comparison of Results GPR could provide a clear picture of the reinforcement bars in both directions, while ultrasonic echo could only reveal the ones in the y-direction, except the one at x = 26 in. This was due to the polarization effects due to the orientation of the ultrasonic echo transducer. The ultrasonic echo could, how- ever, reveal the presence of a localized anomaly. That anomaly could not be detected using GPR. The anomaly appeared between x = 14 in. and x = 30 in. Ultrasonic echo results indicated multiple reflections with changing phase shifts, suggesting a shallow delamination. The IE frequency spectra were disturbed at the location of the anomaly, containing frequency peaks of lower frequencies than that of the backwall

284 echo, also indicating the presence of shallow delamination. However, the nature of the anomaly could not be con- firmed. Both acoustic methods detected the backwall at approximately z = 25 in. Shallow reinforcements were best seen using GPR. Defects indicating a change of impedance from concrete to a material softer than the surrounding tunnel lining could be detected by the acoustic methods: ultrasonic echo and IE. A delamination- like anomaly was indicated. The depth of the anomaly was dif- ficult to estimate, as ultrasonic echo analysis showed multiple reflections of the defect within a cone broadening with depth. IE spectra contained peaks of lower frequencies, indicating a shallow reflector. Test Field 4 Description of the Test Area The last set of measurements at Chesapeake Channel Tun- nel was taken inside the tunnel itself. The testing took place overnight, as traffic control measures were needed. Mea- surements were taken on tile-covered tunnel walls. The tiles were 2 in. by 2 in. Test area CPB4 was located between Station 486+28 and Sta- tion 487, close to the north end of the tunnel, where an anom- aly was previously detected in SPACETEC thermal images. As seen in Figure Q.88, the scanner’s feet were mounted on the wall near the joint. The test field was larger, covering a 48-in. by 36-in. area. The spacing for ultrasonic echo and IE was set at 1 in., and for GPR at 2 in. During the testing, the GPR, ultra- sonic echo, and IE transducers were oriented such that their length was parallel to the centerline of the tunnel. The mea- surement started in the lower right corner of the test field. x = 1.5 in. x = 20 in. Reinforcement y-direction Backwall (a) (b) z y Figure Q.85. CPB3, ultrasonic echo: B-scans showing backwall at z  25 in. and reinforcement in y-direction (a) and through anomaly that caused backwall echo to disappear (b). (b) (a) (c) Figure Q.86. CPB3, US: 3-D image of volume positioned to reveal backwall and its shadowed area due to a shallow anomaly. With B-scan (a) positioned at x  21 in., C-scan (b) at z  25 in., and D-scan (c) at y  12 in.

285 Anomaly A-Scan x = 6 in. A-Scan x = 20 in. (a) (b) (c) f x Figure Q.87. CPB3, IE: D-scan of volume (a) and selected A-scans representing echo from sound areas (b) and from areas with anomalies (c). Figure Q.88. Test area CPB4 near Station 48628.

286 The scanner moved upward and then left (south). Figure Q.88 shows snapshots of the test area and measurement system. GPR Results Figure Q.89, a 3-D image, summarizes the reinforcing elements within the volume. Steel bars were found in both the x- and y-directions in GPR C-scans. The bars in the y-direction appeared to be spaced 12 in. apart, at x = 0 in., x = 12 in., x = 24 in., x = 36 in., and x = 48 in., running from 7-in. to 8-in. deep. The bars in the x-direction were at y = 4 in., y = 12 in., and y = 30 in., at z = 6 in., z = 10 in., and z = 8 in., respectively. Figure Q.90 includes several C-scans at various depths. Two D-scans taken at y = 2.5 in. and y = 12 in. are shown in Figure Q.91. Figure Q.92a is a B-scan showing the rebars in the x-direction cut at x = 18 in. Three rebars at different depths can be distin- guished. Figure Q.92b is a B-scan taken at x = 12 in., showing one bar in the y-direction. The rebar does not run parallel to the surface but is curved, which explains why the rebars are seen at different depths in the C-scans. The rebar curves from z = 8 in. at x = 0 in. up to z = 6 in. at x = 36 in. The rebars in the x-direction run above and under those in the y-direction. Ultrasonic Echo Results The automatic scanning using ultrasonic echo provided no useful information about the condition of the lining at CPB4. Information about the bonding of the tiles could have been gained by analyzing the individual A-scans. However, the ultrasonic echo transducer was too large (4 in. by 3 in.) com- pared with the size of the tiles (2 in. by 2 in.). The grid location and spacing had to be adjusted such that meaningful data (one A-scan per tile) could be obtained. However, the measure- ments were interrupted by an unforeseen weather condition, and no further measurements could be obtained. Impact Echo Results Intensifying foggy weather conditions interrupted the testing during IE data collection because the tunnel had to be fully opened to traffic. Therefore, only four scan lines were taken. (b) (a) (c) Figure Q.89. CPB4, GPR: 3-D image of volume positioned to reveal reinforcement, with B-scan (a) positioned at x  11 in.; C-scan (b) at z  10 in., 4-in. width, and D-scan (c) at y  4 in.

287 z = 1.6 in. z = 6 in. z = 7 in. z = 8 in. z = 10 in. (a) (b) (c) (d) (e) Steel bar x-direction Steel bar x-direction Steel bars y-direction Steel bar x-direction Steel bars y-direction Figure Q.90. CPB4, GPR: C-scans at different depths: (a) z  1.6 in., (b) z  6 in., (c) z  7 in., (d) z  8 in., and (e) z  10 in.

288 Steel bar x-direction Steel bars y-direction y = 2.5 in. y = 12 in. (a) (b) Steel bar x-direction Figure Q.91. CPB4, GPR: D-scans of steel bars along (a) x- and (b) y-directions. Rebars in x-direction run above and under bars in y-direction. Steel bar y-direction Steel bars x-direction x = 18 in. x = 12 in. Figure Q.92. CPB4, GPR: B-scans taken at (a) x  18 in. and (b) x  12 in.

289 Earlier manual measurements indicated that IE is able to evaluate the bonding between tiles and walls when A-scans of individual tiles are analyzed. In the case of automated scan- ning, the analyzed signal is the average of 20 signals recorded close to the source. The dimension of the receiver is 4 in. by 3 in., which covers the area of two tiles. Therefore, evaluating the condition of the bonding of one tile on the basis of auto- matically collected A-scans is not exact. The grid location and spacing had to be adjusted such that meaningful data (one A-scan per tile) could be obtained. However, because of the interruption of the measurements, this was not possible. Figure Q.93 illustrates one of the D-scans along with two representative spectral and temporal A-scans, showing signals/ spectra from apparently debonded and bonded areas. Comparison of Results GPR signals were not disturbed by the presence of the tiles and could capture the reinforcement mesh behind the lining. The IE signals carried useful information about the bonding condition at tile-concrete interface and occasionally about the lining itself. The grid location and spacing had to be adjusted such that meaningful ultrasonic echo data (one A-scan per tile) could be obtained. However, the measurements had to be suddenly stopped. References Colorado DOT. Eisenhower Tunnel. http://www.coloradodot.info/ travel/eisenhower-tunnel. Accessed July 2011. Salek, M. E. 2002. Glenwood Canyon: An I-70 Odyssey. October. http:// www.mesalek.com/colo/glenwood. Schickert, M., M. Krause, and W. Müller. 2003. Ultrasonic Imaging of Concrete Elements Using Reconstruction by Synthetic Aperture Focusing technique. ASCE Journal of Materials in Civil Engineering, Vol. 15, No. 3, pp. 235–246. Shokouhi, P. 2005. Comprehensive Evaluation of Concrete Bridge Decks Using Impact Echo. PhD dissertation. Rutgers, New Brunswick, N.J. Zoega, A., R. Feldmann, and M. Stoppel. 2012. Praktische Anwendun- gen Zerstörungsfreier Prüfungen und Zukunftsaufgaben February 23–24. Fachtagung Bauwerksdiagnose. Berlin: BAM Berlin. debonded fully bonded (a) (c)(b) A-Scan x = 6 in. A-Scan x = 15 in. f x Figure Q.93. CPB4, IE: (a) D-scan, (b) typical A-scans for area with seemingly debonded tiles, and (c) A-scans for area with seemingly bonded tiles.