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.

53 a p p e N D I x I Introduction Since 1982, SPACETEC has offered a scanner system to monitor disruptions and conditions of tunnel linings (Fig- ure I.1). With this tool, it is possible to validate the effects of degradation—such as crack developments, cavities beneath the surface, changes in material composition, and water intrusions—over time. The SPACETEC TS3 scanner is able to record high-precision surface, thermographic, and three-dimensional (3-D) images simultaneously with a resolution of 10,000 pixels and a recording angle of 360° (Figure I.2). The scanner is capable of identifying cracks as small as 0.3 mm in width. The rotating mirror speed of up to 300 Hz is one of the crucial features affecting the measuring speed. It can obtain a fast and nondestructive measurement with only a short period of traffic disruption. The scanner is compact and can be installed in almost every road vehicle that offers enough space for the scanner head and the operator console, such as a minivan. The data are visualized with an easy-to-use, powerful soft- ware package displaying all three channels (visual, thermal, and 3-D) simultaneously. In this way, the operator can inspect the tunnel on the screen, with a pixel-by-pixel synchronism of the recordings. This technique helps the operator analyze and identify suspicious anomalies and compare them on all three channels. Image manipulation is also possible (e.g., adjusting the contrast and brightness of the display, creating 3-D presenta- tions, and performing a 3-D zoom of image details). In many cases, long-term monitoring supports the observation of the tunnel degradation over time with multiple measurements and a recording interval of at least 1 year. A survey of the Chesapeake Tunnel—a part of the Chesa- peake Bay Bridge-Tunnel system in Virginia—was performed in April 2011. The survey was conducted according to the sub- contract agreement with the Federal Institute for Materials Research and Testing (BAM) in Germany. The Chesapeake Bay Bridge-Tunnel is a 37-km-long link crossing the mouth of the Chesapeake Bay and connecting the Delmarva Peninsula’s Eastern Shore of Virginia with Virginia Beach and the metropolitan area of Hampton Roads. The bridge-tunnel system combines bridges—connecting four artificial islands—with the Timble Shoal Tunnel (the western side of the bay) and the Chesapeake Tunnel. The Chesapeake Tunnel is one of two immersed, sunken-tube tunnels con- structed under the ship channels of Chesapeake Bay in an approximately east–west direction and was opened in 1964. Since it opened, the tunnel has been exposed to extreme envi- ronmental conditions. Water intrusion and corrosion have been reported during visual inspections. The portal-to-portal length of the tunnel is 1,661 m, with a roadway (two-lane) width of 7.3 m plus a sidewalk on one side. The tunnel interior is faced with ceramic tiles, which is uncommon for non-U.S. tunnels. Therefore, the surface of the concrete lining underneath is inaccessible. This appendix describes the methods and results of the survey and is divided into the following parts: • Available data channels; • The recording process, including scanner measurements and scanning parameters; • A description of results, including a brief introduction to data processing and a detailed description of the data; and • Working with the data. The corresponding data sets, including the analysis soft- ware package, were shipped with an external hard drive to BAM on January 6, 2011. available Data Channels Visual Images Visual images (Figure I.3) are most frequently used for gen- eral documentation and maintenance purposes. They show the condition of the lining as far as visible phenomena are Survey of the Chesapeake Bay Bridge-Tunnel

54 Figure I.3. Visual image of conditions of a shotcrete lining in a motorway tunnel. Figure I.1. SPACETEC TS3 scanner. Figure I.2. Scanning principle.

55 concerned. At any time, later data can be consulted to look for changes in those conditions. Profile Data Profile data (Figures I.4 and I.5) show the dimensions of the tubes and are used to consider and solve clearance problems. In the small range, the presence of distance measurements at high density allows the inspector to find and identify surface defects (e.g., a chip-off or spalling) in the lining. Thermal imaging (Figure I.6) measures the surface tem- perature in the tunnel interior. Temperature differences determine information about the state of the lining. The differences can result from various interacting processes between the surface of the lining and the air in the tunnel, such as the following: • Cooling as a result of water evaporating from the surface; • The reaction of the lining material during cooling or heating; Figure I.4. Structural gauge investigation in motorway tunnel. Red spots indicate obstructions to given clearance profile. Users can determine cut volume and affected area on screen. Figure I.5. 3-D view of concrete surface in tunnel with damage (chip-off) near the joint of two sections.

56 • The influence of cold and warm temperatures, respectively, at the surface; • Surface roughness; • Cavities (gravel nests below the surface, bad contact of the lining to the rock, and gravel rock material); and • Nonhomogeneous material composition. Using detailed, known measuring conditions as a compari- son, the user can exclude certain thermal interactions and ensure a correct interpretation of the thermal imaging. A quasi- stationary heat flow between the air in the tunnel and the rock behind the lining creates suitable measuring conditions. If the heat-flow conditions are unknown, certain features cannot be clearly identified. Still, the thermographic image displays sig- nals that offer supplemental information to the visual image (see Figure I.6). This information can be used to highlight some effects, like water evaporation. The premeasurement program used in this project was part of the thermographic survey and evinced proper conditions for the recording. The temperature survey had to be done in a short period because of the constantly changing temperatures in the tunnel. The SPACETEC TS3 scanner system was able to perform such a fast and reliable measurement. recording process Table I.1 summarizes the data summary from the Chesapeake Tunnel. With the inspection vehicle used for this survey (Fig- ure I.7), the intended driving speed of 1.5 km/h could not always be kept constant (the speed went as high as 3.5 km/h). Therefore, some pixels are stretched in the driving direction. Driving too fast may cause gaps in the laser scan lines at the tunnel wall, which can influence the visibility of cracks. To Figure I.6. Example of thermal image showing water infiltration. Water evaporation on the surface yields a clear cold signal. This method marks even smaller water spots clearly.

57 Table I.1. Data Summary of the Chesapeake Tunnel Time of measurement April 11–12, 2011 Scanning length 1,680 m Vehicle speed Approximately 1.5 km/h Recording channel Infrared (8–12 µm) 10,000 px/scan Visual 10,000 px/scan Profile 10,000 px/scan Mirror speed 160 Hz Temperature resolution Approximately 0.1°C Spatial resolution 3 mm by 3 mm at the surface Figure I.7. Inspection vehicle in Chesapeake Tunnel. avoid such problems, a nearly constant speed should be maintained during the survey. The survey was performed during the night spanning April 11 and 12, 2011. During the survey, a sufficient temperature dif- ference for a quasistationary heat flow was obtained. The TS3 scanner was installed on the roof at the rear of the inspection vehicle (Figure I.7). That placement provided an un disturbed 360° measurement. The highest resolution of 10,000 pixels was used to obtain an appropriate imaging of fine-scale features. A complete traffic closure was not possible. Thus, the record- ing was performed twice: once in the north–south direction of the lane to Virginia Beach and once in the opposite lane (south– north) toward the eastern shore of Virginia. Traffic could pass the inspection vehicle, as is visible in the recordings. Description of results Data Processing The recorded data were corrected for geometry, and the 360° display of the tunnel was projected with a defined scale onto a plane surface (Figure I.8) for a synchronous display of all three channels: visual, thermal, and 3-D. One lane was recorded in the north–south driving direc- tion, and the second lane was recorded in the opposite (south–north) direction. Figure I.8 is labeled with the cor- responding driving direction (south or north) and with an absolute true-scale location in meters. Common artificial installations like hand rails, air ports, and electrical and maintenance installations are highlighted. They are clearly visible in both data channels. The cement conduits behind the ceramic tiles are only visible in the thermal image and correlate with information from the construction plans of the tunnel. A full data set comprised visual, thermal, and 3-D channels and was formatted and edited to evince a true-to-scale display, labeled with a meter range (a change in feet was also possible, if needed). The thermal data were corrected by the commonly existing air temperature drift along the tunnel axis. After level- ing, thermal data were displayed with a constant air tempera- ture. Therefore, the same phenomena were displayed with the same colors. The data interpretation was based on local tem- perature differences (anomalies); thus, an absolute tempera- ture was not needed. Every thermographic surface point corresponded to a color-coded temperature interval with a temperature resolu- tion of 0.1°C (Figure I.9) and 16 colors ranging from black through blue, green, red, and yellow to white. This color pal- ette gave an intuitive physiological impression of cold (dark to blue) and warm (red to white) temperatures. Figure I.8 highlights the most common installations in the dataset, which could mainly be ascribed to artificial origins, such as the following: • Fresh and exhaust air ports and corresponding swirled air; • Hand railings, niches, and supply boxes installed in the lining wall; • Traffic lightings and signs; and • Tubes behind the lining, visible in the thermal image. Visual Results and Distance Measurements Since opening in 1964, the Chesapeake Tunnel has been exposed to strong environmental effects such as exhaust gases from traffic, corrosion, and water intrusion. In general, the visual damages are easy to identify and are self-explanatory. Fig- ure I.10 displays split and dirty ceramic tiles, which are the main concern. Not only are profiles important for clearance consider- ations in the railway sector, the distance measurements can also be used to characterize damage. The dimensions of the damaged areas can be easily worked out.

58 S N left ceiling right driving direction NS right left ceiling exhaust air port fresh air port cement conduits for power cables hand rail tunnel lighting driving direction Projection to a plane lane fire extinguisher niche electrical installation fire extinguisher niche & road sign swirled warm air Figure I.8. Perspective view (bottom) demonstrates projection of cylindrical-shaped tunnel onto a plane (top). Top panel displays true-scale projected tunnel with a visual (left) and thermal (right) channel from an interior view. Both synchronized channels show same location and same content with different data sets. Figure I.9. Color scale for thermographic images.

59 Figure I.11 shows one common type of damage—broken or missing tiles—which is clearly evident in the visual chan- nel (upper left). A perspective view of the area (A, lower left) displays more details and is useful for damage assessment. A distance profile (B, lower right) helps estimate the dimen- sions of the disruptions or highlight artificial installations like emergency lighting. Thermal images usually show a clear cut in the outer rim of disruptions. In this case (upper right), the right side is positioned in the wind shadow and is cooler (darker blue), and the left side is exposed to the warm, lighter air flow coming from the right side of the image. In the area directly ahead and 90° to the side of the detector, the sensor is overmodulated, and the intensities of the reflected signal are high, which is highlighted by an intense horizontal stripe in both channels. The loose and broken tiles are mainly located on the ceiling of the tunnel (Figures I.10 and I.11). Classification of Thermal Anomalies Thermal images consist of a thermal conduction from the tunnel interior into the rock. This determines the qualitative correspondence to the nature of the heat source, as shown in Table I.2. Some local temperature anomalies can be explained by construction factors, for example, air swirls resulting from obstacles (road signs and traffic lights), niches (which can be recognized in visual images), and tubes behind the linings. Detailed analysis and interpretation of the data were applied interactively on the screen. Visual and thermal images were analyzed simultaneously to figure out some correspon- dences between temperature-related patterns and visible construction. The color-coded temperatures and the color resolution were adjusted to the specific temperature anomaly to improve the visibility of the objects. The Premeasurement Program A pretesting unit was installed in a fire extinguisher niche 250 m from the western portal (Virginia Beach). The unit could not be installed at the place with the worst-case conditions for thermal measurements in the middle of the tunnel because the distance for data transfer through a cable to the next telephone plug would have been too long. The premeasurement program was used before and during the recording of the thermal image. It allowed for advanced determination of the time and weather conditions that would be favorable for the purpose of the survey. It documented the Figure I.10. Development of split ceramic tiles: south, 544 m (left); north, 163 m (right).

60 A A B B Figure I.11. Views of broken or missing tiles.

61 required heat flow conditions during the thermal measure- ment between the lining and the rock to resolve and interpret patterns of heat anomalies. Temperature sensors were placed in the target structure. One sensor measured the air temperature, the second sensor measured the material temperature near the surface at a depth of 0.075 m, and the third sensor measured deeper depths of 0.3 m. The data logger (master) read and permanently stored the temperature recordings (usually one record per hour) from the sensors (Table I.3). The data could be accessed via telephone line and displayed on the screen (Figure I.12). Thermal Results The temperature at the tunnel surface reflected the heat con- duction of the lining below the surface. Figure I.13 displays an example of a heat flow under different material conditions with a cavity or wet spots (which are not visible at the surface). The displayed situation is typical for warmer seasons: the air temperature in the tunnel is higher than the rock temperature. The stationary heat flow between the air temperature and the rock resulted in a surface temperature that depended on the heat conductivity of the lining. A cavity reduced the heat con- ductivity and resulted in a higher surface temperature. There- fore, the tunnel thermography revealed damages in the lining when those damages influenced conductivity. The quasistationary measuring conditions were adjusted naturally with the corresponding weather conditions and air Table I.2. Dependency Between Temperature Anomaly and Heat Conduction Anomaly Thermal Conduction Possible Reasons Cold Better Good thermal contact between rock and lining: • Water in lining • Higher density of the material Warm Worse Bad thermal contact between rock and lining: • Loose, less lithified rock • Lower density of the material • Higher porosity, hollow spaces Table I.3. Premeasurement Program in the Chesapeake Tunnel Location in the Tunnel Sensor Depth Remarks 250 m east of the Virginia Beach portal In front of the lining Air temperature 0.075 m into the lining Temperature difference 0.3 m into the lining Temperature difference Figure I.12. Temperature alignment before thermographic survey. (Note: stunde = hour.) measurement 11 Apr. 12 Apr.

62 temperature when the tunnel had proper air convection— resulting from a chimney effect, caused by different air pressures between the tunnel portals, or from steady traffic. Long-term surveys of other tunnels revealed a number of good measuring conditions during a period of several months. Figure I.14 displays the effect of the thermal reflections of the installed constructions on the bended corners of the tube. The ceramic tiles seemed to have a higher reflectivity in the infrared spectrum. This is uncommon for concrete or brick- work tunnel linings. Figure I.15 shows the direction of flow of warmer air from the port in the tunnel ceiling. Figure I.16 reveals some linear structures behind the lining of a side wall, which is referred to as drainage channels with lower heat conductivity and higher temperatures. This indi- cates that the building documentation needs to be reviewed before further investigations. In the overall length of the tunnel, several temperature- related anomalies were detected (Figure I.17). Some of the larger temperature anomalies could not be ascribed to artificial sources. The thermography displayed a center with lower tem- peratures (higher heat conductivity) surrounded by a rim of higher temperatures (lower heat conductivity). The origin of these anomalies was unknown. Figure I.18 displays a common feature that was visible in both the visual and the thermal Figure I.14. Thermal shadows of installation. Ceramic tiles have higher reflectivity in infrared spectrum (north, around 825 m). electrical installationniche lost tile thermal shadows Figure I.13. Dependency of surface temperature on heat conduction of lining material. Heat flow lining airrock normal porous wet

63 Figure I.15. Air flow from port in tunnel ceiling. Warmer air flows in right-side direction, clearly visible in thermal image (north, 1,654 m). Figure I.16. Visible structures at the surface and drainage behind lining surface (south, 1,038 m). Drainage behind the lining dirt metal plate

64 defective and loose tiles anomalies Figure I.17. Loose tiles (left) and warmer temperature anomalies (right) in ceiling area (south, 636–652 m). Figure I.18. Renewed tiles with different, compacted material (south, around 675 m).

65 datasets: joints between the tiles showed a different reflectivity, which seemed to indicate renewed ceramic tiles. The main findings in the thermographic data set were these: • Cable channels and drainage tubes behind the linings with lower heat conduction; • Areas with lighter tile joints, maybe renewed or repaired tiles, with different materials and lower heat flow at the side walls; and • Areas with larger anomalies behind the ceiling walls. The thermographic images did not always make clear exactly what was behind a surface. Therefore, further investi- gation was necessary to determine the reasons for the weak points in the lining. Working with the Data The software package TuView was the tool used to analyze and display the data sets of the corresponding three chan- nels. Data access was provided by information files con- taining the specifications for the image files as well as the true-scale information. TuView offered the ability to highlight zones of interest with different color codes. The information was saved in notebook files, which were delivered with the report. The notebook files were separated into the following categories: blue, indicating artificial installations like road signs or traffic lights; red, indicating damaged areas (loose and bro- ken tiles); and green, indicating anomalies of unknown origin.