Concrete Pavement Joint Diagnostics with Ultrasonic Tomography

Kyle Hoegh, Lev Khazanovich, and H. Thomas Yu

An important issue in extending the life of jointed plain concrete pavement is the proper construction and maintenance of transverse joints. The capabilities of an ultrasonic tomography device in providing information essential to joint rehabilitation decisions as well as the quality control of rehabilitation construction are presented. Testing was conducted at various partial-depth repair locations of a recently repaired in-service highway to determine the presence and extent of debonding between the partial-depth repair and the existing concrete. Testing also was conducted at the Minnesota Road Research Facility to evaluate the subsurface condition along various transverse joints. Real-time diagnosis with the intensity-based signals of the synthetic aperture focusing technique was successful in pinpointing the location of debonded concrete interfaces within partial-depth repairs. Ultrasonic tomography testing also showed the capability of identifying that a joint is in need of repair and subsequently diagnosing the extent of subsurface deterioration throughout the joint.

Transverse joints are saw cut into jointed plain concrete pavement to control the cracking caused by expansion and contraction of the portland cement concrete (PCC) slab in response to temperature and moisture variations in the concrete layer. Even though dowels are used to improve the performance of such joints by ensuring proper load transfer between the two slabs, joint introduction can cause pavement distresses as a result of environmental and wheel loading. Rehabilitation strategies such as partial-depth repair can be used to mitigate or repair joint distresses such as spalling at the surface. Visual inspection is an effective method of determining the surface distress of a transverse joint. However, the distress cannot be identied until it manifests at the surface, and the extent of subsurface damage cannot be determined. Distress detection before surface exposure can improve rehabilitation timing, and determination of the extent of subsurface damage can assist in the choice of the appropriate rehabilitation type (i.e., partial- versus full-depth repair). Nondestructive evaluation of the quality of construction or repair at a joint (e.g., verifying a full bond between the repair materials and existing concrete in a partial-depth repair) can aid decisions about construction or reconstruction. Because these issues involving PCC joint evaluation are important to increasing the service life of the pavement, an effective nondestructive method for this application would be of signicant value.
K. Hoegh and L. Khazanovich, Department of Civil Engineering, University of Minnesota, 560 Pillsbury Drive SE, Minneapolis, MN 55455. H.T. Yu, FHWA, HIPT30, Room E73-444, 1200 New Jersey Avenue SE, Washington, DC 20590. Corresponding author: L. Khazanovich, khaza001@umn.edu. Transportation Research Record: Journal of the Transportation Research Board, No. 2305, Transportation Research Board of the National Academies, Washington, D.C., 2012, pp. 5461. DOI: 10.3141/2305-06 54

Nondestructive deection testing techniques that have been used to evaluate concrete pavement joints include ground-penetrating radar, magnetic pulse induction, chaining, and other seismic methods, such as impact echo. Past studies have shown that groundpenetrating radar can obtain high-speed pavement measurements, but accuracy decreases with increased speed, and the detection of nonuniform cracks is difcult (16). Magnetic pulse induction can locate metal inclusions with high accuracy, but it does not detect any type of nonmetallic inclusion or defect (710). Infrared ther mo graphy can indicate subsurface anomalies such as delamination in concrete, but this method is limited by environmental conditions and depth of the defect (1113). Impact echo has been shown to detect thickness and planar aws, but nonplanar aws that do not result in resonance can be difcult to detect (1416). Chain dragging is a method that state agencies use extensively for PCC joint evaluation, among other applications. This method can be effective when the operator is experienced and skilled, but the analysis is qualitative and does not result in a data record that can be analyzed further. This paper explores the use of an ultrasonic tomography device commercially known as MIRA for the evaluation of concrete pavement joints. High-frequency (>20,000-Hz) sound waves generated by transducers are used to characterize material properties or detect their defects. In the pitch-catch method used in this study, the sound waves are guided into the PCC and the resulting reections analyzed. The dry point contact transducers used in this study transmit lowfrequency (55-kHz) shear waves that penetrate to the depths necessary to evaluate the subsurface of a concrete structure (17, 18). Such transducers have been used for years for the detailed evaluation of civil structures such as tendon ducts and bridge decks (1922). Figure 1 (left portion) illustrates a manual MIRA measurement that uses the 45 transmitting-and-receiving pairs from the 40-transducer, 10-channel ultrasonic array and the synthetic aperture focusing technique (SAFT) to create a two-dimensional reconstruction of the concrete directly below where the scan is taken (SAFT B-scan) in approximately 1 s (23). Also in Figure 1, the 45 measurement pairs used in each MIRA scan provide an increased redundancy of information of MIRA (shown at bottom right) over conventional impact echo (shown at top right). MEASUREmENT PROCESS AND SIGNAL INTERpRETATION Each of the 45 transmitting-and-receiving pairs from the array of transducer channels in each MIRA scan gives information about intensity versus time. To visualize the subsurface condition below each measurement, the time-of-ight arrivals must be converted to distance. The shear wave velocity used to make this conversion is measured in calibration mode by dividing the xed distance between the transducer pairs by the respective shear horizontal arrival times.

Hoegh, Khazanovich, and Yu

55

Impact

Receiver

-8

Local Transverse Position, in.

Impact Echo

1 pair per measurement

MIRA

45 pairs per measurement

Depth, in.

Backwall reflection (~9 in.)

FIGURE 1 MIRA ultrasonic pitch-catch device and comparison with traditional impact echo method ( 15, 23 ) (1 in. 5 25 mm).

12

SAFT uses this characteristic shear horizontal wave velocity in scan mode to reconstruct the medium below the measurement on the basis of the body shear wave reections that arrive at the surface. SAFT has been found to be a feasible algorithm for use with ultrasonic pitch-catch technology as well as other applications (21, 23). Each approximately 1-s MIRA scan gives a two-dimensional depth cross section (SAFT B-scan) with the vertical axis indicating the depth of any reection (caused by change in acoustic impedance) and the horizontal axis indicating the location along the aperture of the device; zero is the center of the scan location. When SAFT B-scans are evaluated, changes in acoustic impedance (affected by material stiffness changes) will cause a high-intensity reection (red) associated with the location of the anomaly. This method can be used to detect debonding between concrete layers, among other subsurface anomalies. Figure 2 is a SAFT B-scan that resulted from a MIRA scan taken at a full-depth repair location. The gure indicates little to no reection (blue) until a high-intensity

FIGURE 3 Example B-scan taken at partial-depth repair location where no metal reinforcements are present (1 in. 5 25 mm).

-8

Local Transverse Position, in.

Backwall reflection (~9 in.) 12

FIGURE 2 Example B-scan taken at full-depth repair area showing typical backwall reflection (red) at interface between base concrete and base at 9 in. (~225 mm).

reection (red) at a depth of approximately 9 in. (225 mm), at the interface between the concrete and base material. This type of strong backwall reection at the depth of the concrete and base interface should be expected if a single concrete layer has no signicant aw. This signal is much cleaner (blue with little to no green, yellow, or red at depths above the backwall reection) than that of even a properly bonded partial-depth repair over existing concrete. Figure 3 is a B-scan taken at a properly bonded partial-depth repair location showing a typical backwall reection at the interface between the concrete and base as well as some additional lowerintensity reection at shallower depths where the new and old concrete interface is located. These lower-intensity reections are caused by the presence of two concrete layers with slightly different acoustic impedances as well as coarse aggregate. The presence of the layers was determined through the consistent appearance of moderate intensity reections in B-scans at the depth of the repair material layer, even in cases where cores veried a proper bond. Metal inclusions also can cause reections of the shear waves. Figure 4 is a scan taken at a transverse joint that indicates the presence of dowels. High-intensity reections near 4 in. (100 mm) indicate the presence of dowels. However, because the MIRA B-scan uses multiple measurement pairs, a backwall reection is still visible between the concrete and base. The SAFT B-scans in Figures 2 through 4 indicate a sound concrete condition with high-intensity reections associated only with locations of metal dowels (if present) and depth of the concrete base interface. Variation from these types of scans can indicate subsurface damage. One type of variation from a typical scan that indicates subsurface damage is called shadowing, which refers to the absence of a high-intensity reection at a location where there is a change in acoustic impedance. The depth of the interface between the top concrete layer and base primarily is used for the shadowing analysis, and this interface is referred to as the backwall in this study. Because of the large difference in acoustic impedance between the concrete layer and the base layer, a high-intensity reection should

Depth, in.

56

Transportation Research Record 2305

-8

Local Transverse Position, in.

intensity reection is observed at the backwall depth in a SAFT B-scan. This type of analysis is useful for diagnosing the presence and extent of subsurface damage at concrete pavement joints, evaluating the quality of rehabilitation construction, or both, as detailed in the section on eld trials. FIELD TRIALS Partial-Depth Repair Minnesota Department of Transportation (DOT) personnel allowed for access to an in-service highway that was recently rehabilitated at the transverse joints with partial-depth repairs. Chaining (the traditional nondestructive deection testing method for this type of diagnostics) was carried out by experienced Minnesota DOT personnel to identify general areas with debonding potential. MIRA ultrasonic tomography scans also were taken where the chaining was performed. The information gathered by both methods was used to determine where to take cores. An example partial-depth repair section where verication cores were taken compares the capabilities of MIRA and chaining for the quality assurance of partial-depth repair bond with existing concrete. Figure 5 shows a series of MIRA scans (Positions 1 through 12) taken at a partial-depth repair (2 ft 2 ft) located at a transverse joint. The left side of this partial-depth repair is approximately 3 ft from the fog line of the truck lane. Each of these scans resulted in real-time SAFT B-scan results associated with the condition of the concrete below (Figure 6).

dowels Depth, in.

Backwall reflection(~9 in.) 12 FIGURE 4 SAFT B-scan taken at partial-depth repair near transverse joint where dowels are present (1 in. 5 25 mm).

be expected at this planar depth. However, if an obstruction exists at a shallower depth, shear waves will either be reected back to the surface or attenuate before they penetrate to the depth of the backwall. Therefore, because each measurement has 45 transmittingand-receiving pairs, a planar obstruction (e.g., debonded partialdepth repair or deteriorated concrete) can be assumed if a low-

Position 1

Position 2

Position 3

Position 6

Position 5

Position 4

Position 7

Position 8

Position 9

Position 12

Position 11

Position 10

FIGURE 5 MIRA scan locations taken at the example of partial-depth repair.

Hoegh, Khazanovich, and Yu

57

Position 1

Position 2 (a)

Position 3

Position 6

Position 5 (b)

Position 4

12

Position 7

Position 8 (c)

Position 9

Position 12

Position 11 (d)

Position 10

FIGURE 6 SAFT B-scans associated with 12 MIRA scan positions (depth of each scan 5 300 mm, 1 in. 5 25 mm): ( a ) Transverse Positions 13, ( b ) Transverse Positions 46, ( c ) Transverse Positions 79, and ( d ) Transverse Positions 1012.

a proper bond between the partial-depth repair and the existing concrete (Figure 7b). SAFT B-scans within the area deemed to be experiencing debonding were matched with their corresponding positions. Additional inspection revealed that a few of the scans in this vicinity did in fact have a strong backwall reection, indicating a proper bond. These scans included the location where the core was taken (Positions 1 and 6). The scan locations and B-scans for Positions 1 and 6 are shown in Figure 8, a and b, respectively. These scans are similar to the scan in Figure 3 that indicates a proper partial-depth repair bond. The clear backwall reection at the interface of concrete and base indicates a proper partial-depth repair bond above. Locations where the backwall reection exhibited clear shadowing (Positions 8 through 11) were inspected further with MIRA measurements to pinpoint a location where the core would verify the debonding within this general vicinity (Positions 9 and 10). The MIRA scan locations and associated B-scans for Positions 9 and 10 are shown in Figure 9, a and b, respectively. These SAFT B-scans are different from the type of SAFT B-scan that would be expected for a properly bonded partial-depth repair. Backwall reection is not present at the interface between the existing concrete and base layer, thus indicating that the unbounded condition at the interface of the partial-depth repair and existing concrete obstructed the shear waves from transmission through to the depth of the concrete. The core taken at this pinpointed location is shown in Figure10b to verify the nondestructive diagnosis of an improperly bonded partial-depth repair. The partial-depth repair is debonded, especially compared with a core from a properly bonded location (Figure10a). A similar process was used to pinpoint a debonded location in other joint locations where MIRA indicated debonding. Figure 11 shows another core that exhibits debonding of the partial-depth repair, as pinpointed by MIRA diagnosis.

12 Depth, in. 0

Depth, in. 0

12 Depth, in. 0

12 Depth, in. 0

Most of these scans exhibit shadowing of the backwall reection at about 9 in. (225 mm), indicating the presence of a planar aw such as debonding, which was corroborated with chaining information for this general vicinity. A core was taken at a location within this partial-depth repair (toward the fog line and away from the transverse joint; Figure 7a), and inspection of the cross section showed

Joint Subsurface Evaluation A eld trial at the Minnesota Road Research Facility was conducted to determine the subsurface condition near transverse joints in terms of presence and extent of potential aws.

Core location (a) FIGURE 7 Partial-depth repair: ( a ) core location and ( b ) cross section showing proper bond. (b)

58

Transportation Research Record 2305

-8

-8

-8 Local Transverse Position, in. 8 Pos1

-8 Local Transverse Position, in. Pos6

Depth, in.

12

Backwall reflection at depth of interface between concrete and base (~9 in.)

FIGURE 8 SAFT B-scans from ( a ) Position 1 and ( b ) Position 6 in location where first core was taken (a strong backwall reflection indicates a proper bond between existing concrete and partial-depth repair) (1 in. 5 25 mm).

-8

-8

-8 0

Local Transverse Position, in.

-8

Local Transverse Position, in. Pos10

Pos9
Depth, in.

12

Shadowing of where the backwall reflection should occur between concrete and base(~9in.)

FIGURE 9 SAFT B-scan locations indicating poor bond between partial-depth repair and existing concrete from ( a ) Position 9 and ( b ) Position 10 (1 in. 5 25 mm).

Hoegh, Khazanovich, and Yu

59

(a)

(b)

FIGURE 10 Comparison of cores: ( a ) properly bonded repair and ( b ) improperly bonded partial-depth repair toward centerline near transverse joint.

Joint Degradation Multiple MIRA scans were taken at various transverse joints in an initial screening. The SAFT B-scans analyzed at one joint showed some locations that exhibited potential aws. More detailed testing along the joint with scans taken in 3-in. step sizes showed the extent of the potential damage at the joint. Analysis of the SAFT B-scans taken in the general area of the right wheelpath showed results similar to those in Figure 4. A SAFT B-scan at a right wheelpath is shown in Figure 12a. In this location and other locations around the right wheelpath, the areas of high-intensity reection correspond to dowel locations, and the depth of the concretebase interface is as should be expected for undamaged concrete. Analysis of the remaining portion of the joint indicated a very different subsurface condition. Subsequent measurements at the center of the joint and moving toward the centerline of the joint indicated damage. Instead of reections strictly at the dowels and concrete thickness, an uneven reection is apparent at a shallower depth that is unexpected. This shallow, uneven reection indicates concrete deterioration at the interface with the base.

Minnesota DOT personnel subsequently took a full-depth concrete sample [12 ft 3 ft (3.658 m 0.914 m)] from along the transverse joint to verify the MIRA SAFT B-scan diagnosis. As shown in Figure 13, the sample was ipped so that the interface between the concrete and base was at the top. The underside of the concrete near the right wheelpath is in relatively good condition, as was diagnosed by MIRA. The underside of the concrete near the base in the samples at the middle of the joint and at the centerline shows signicant deterioration. This examination veried the nondestructive diagnosis of deterioration with SAFT B-scans before forensic verication. CONCLUSIONS The testing presented in this study indicated that although chaining with experienced personnel can identify the general qualitative information, MIRA testing and analysis is an attractive alternative to improve diagnostic capabilities. With the analysis of SAFT B-scan real-time signals, ultrasonic tomography identied the presence of debonding, pinpointed details regarding the precise location (boundaries) of the debonding, reduced the diagnostic subjectivity, and archived a quantitative record of scan locations and results to allow for additional analysis. Ultrasonic tomography testing also was conducted to diagnose the subsurface condition at various transverse joints. Analysis of SAFT B-scans and subsequent forensic analysis showed MIRA to be capable of identifying subsurface distress before manifestation at the surface. MIRA also diagnosed the location and severity of deterioration in the concrete subsurface. Even though MIRA is an attractive tool for the evaluation of the condition of joints and partial-depth repair bonds, some steps can be taken to make the measurement process and evaluation more user friendly. A provisional specication for MIRA testing of transverse joints and partial-depth repair quality control would be benecial in making the measurement process and signal interpretation more productive for routine diagnostics.

FIGURE 11 Core showing debonding of partial-depth repair.

60

Transportation Research Record 2305

-8

Local Transverse Position, in. Dowel

-8

Local Transverse Position, in. Deteriorated Concrete

Depth, in.

10 Backwall Reflection

20 (a) (b)

FIGURE 12 MIRA B-scan locations in suspected ( a ) sound and ( b ) deteriorating conditions (1 in. 5 25 mm).

(a)

(c)

(b) FIGURE 13 Forensic sample sections used to verify concrete condition on underside at various locations near transverse joint: ( a ) centerline, ( b ) middle of joint, and ( c ) right wheelpath.

ACKNOWLEDGMENTS The authors thank Tom Burnham and Tim Clyne of the Minnesota Department of Transportation (DOT) for supplying eld testing sections from the Minnesota Road Research Facility for analysis. The authors acknowledge Maria Masten and Gordy Bruhn of the Minnesota DOT for providing access to the partial-depth repair testing and conducting the chain dragging. Shongtao Dai of the Minnesota DOT was the technical liaison of the project under which this study took place.

REFERENCES
1. Gucunski, N., F. Romero, S. Kruschwitz, R. Feldmann, A. Abu-Hawash, and M. Dunn. Multiple Complementary Nondestructive Evaluation Technologies for Condition Assessment of Concrete Bridge Decks. In Transportation Research Record: Journal of the Transportation

Research Board, No. 2201, Transportation Research Board of the National Academies, Washington, D.C., 2010, pp. 3444. 2. Gucunski, N., C. Rascoe, R. Parrillo, and R.L. Roberts. Complementary Condition Assessment of Bridge Decks by High-Frequency GroundPenetrating Radar and Impact Echo. Presented at 88th Annual Meeting of the Transportation Research Board, Washington, D.C., 2009. 3. Abdallah, I.N., R. Williams, and S. Nazarian. Application of Data Fusion Using Fuzzy Logic Method to Nondestructive Evaluation of Pavements. Presented at 88th Annual Meeting of the Transportation Research Board, Washington, D.C., 2009. 4. Li, J., D.G. Zollinger, and R.L. Lytton. Detection of Delamination in Concrete Pavements Using Ground-Coupled Ground-Penetrating Radar Technique. In Transportation Research Record: Journal of the Transportation Research Board, No. 2087, Transportation Research Board of the National Academies, Washington, D.C., 2008, pp. 6877. 5. Maierhofer, C. Nondestructive Evaluation of Concrete Infrastructure with Ground Penetrating Radar. Journal of Materials in Civil Engineering, Vol. 15, No. 3, 2003, pp. 287297. 6. Chong, K.P., N.J. Carino, and G. Washer. Health Monitoring of Civil Infrastructures. Smart Materials and Structures, Vol. 12, No. 3, 2003, pp. 483493.

Hoegh, Khazanovich, and Yu

61

7. Rao, S., K. Hoegh, H.T. Yu, and L. Khazanovich. Evaluation of Dowel Alignment Constructability Evaluation in Portland Cement Concrete Pavements. In Transportation Research Record: Journal of the Transportation Research Board, No. 2098, Transportation Research Board of the National Academies, Washington, D.C., 2009, pp. 8693. 8. Hoegh, K., T. Yu, and L. Khazanovich. MIT SCAN Application in North America: Five Years After. Proc., 9th International Conference on Concrete Pavements, San Francisco, Calif., 2008, pp. 450462. 9. Hossain, S., and M. K. Elno. Field Demonstration of Magnetic Tomography Technology for Determination of Dowel Bar Position in Concrete Pavement. VTRC 06-R40. Virginia Transportation Research Council, Charlottesville, Va., June 2006. 10. Khazanovich, L., H.T. Yu, and R. Stubstad. Nondestructive Dowel Bar Detection in Existing Rigid Concrete Pavement Slabs. California Department of Transportation, Sacramento, 2003. 11. Washer, G.A., R. Fenwick, N. Bolleni, and J. Harper. Effects of Environmental Variables on Infrared Imaging of Subsurface Features of Concrete Bridges. In Transportation Research Record: Journal of the Transportation Research Board, No. 2108, Transportation Research Board of the National Academies, Washington, D.C., 2009, pp. 107114. 12. Clark, M.R., D.M. McCann, and M.C. Forde. Application of Infrared Thermography to the Non-Destructive Testing of Concrete and Masonry Bridges. NDT and E International, Vol. 36, No. 4, 2003, pp. 265275. 13. Maser, K.R., and W.M.K. Roddis. Principles of Thermography and Radar for Bridge Deck Assessment. ASCE Journal of Transportation Engineering, Vol. 116, No. 5, 1990, pp. 583601. 14. Schubert, F., and B. Kohler. Ten Lectures on Impact-Echo. Journal of Nondestructive Evaluation, Vol. 27, 2008, pp. 521. 15. Carino, N.J. The Impact-Echo Method: An Overview. Proc., 2001 Structures Congress and Exposition (P.C. Chang, ed.), May 2123, 2001, Washington, D.C., American Society of Civil Engineers, Reston, Va., 2001, pp. 118.

16. Gibson, A., and J.S. Popovics. Lamb Wave Basis for Impact-Echo Method Analysis. ASCE Journal of Engineering Mechanics, Vol. 131, No. 4, 2005, pp. 438443. 17. Nesvijski, E.G. On the Problem of Application of the Conic and Exponential Wave Guiding Extensions for Ultrasonic Transducers for Materials Testing. Nasta Technical Bulletin, Vol. 3, 1997, pp. 4956. 18. Mayer, K., K.-J. Langenberg, M. Krause, B. Milmann, and F. Mielentz. Characterization of Reector Types by Phase-Sensitive Ultrasonic Data Processing and Imaging. Journal of Nondestructive Evaluation, Vol. 27, Nos. 13, 2008, pp. 3545. 19. Schubert, F., and B. Khler. Three-Dimensional Time Domain Modeling of Ultrasonic Wave Propagation. Journal of Computational Acoustics, Vol. 9, No. 4, 2001, pp. 15431560. 20. Khazanovich, L., R. Velasquez, and E.G. Nesvijski. Evaluation of Top-Down Cracks in Asphalt Pavements by Using a Self-Calibrating Ultrasonic Technique. In Transportation Research Record: Journal of the Transportation Research Board, No. 1940, Transportation Research Board of the National Academies, Washington, D.C., 2005, pp. 6368. 21. Langenberg, K.J., M. Brandfa, R. Hannemann, C. Hofmann, T. Kaczorowski, J. Kostka, R. Marklein, K. Mayer, and A. Pitsch. Inverse Scattering with Acoustic, Electromagnetic, and Elastic Waves as Applied in Nondestructive Evaluation. In Waveeld Inversion (A. Wirgin, ed.). Springer, New York, 1999, pp. 59118. 22. Marklein, R., K. Mayer, R. Hannemann, T. Krylow, K. Balasubramanian, K.J. Langenberg, and V. Schmit. Linear and Nonlinear Inversion Algorithms Applied in Nondestructive Evaluation. Inverse Problems, Vol. 18, No. 6, 2002, pp. 17331759. 23. Hoegh, K., L. Khazanovich, and H.T. Yu. Ultrasonic Tomography Technique for Evaluation of Concrete Pavements. In Transportation Research Record: Journal of the Transportation Research Board, No. 2232, Transportation Research Board of the National Academies, Washington, D.C., 2011, pp. 8594.
The Rigid Pavement Design Committee peer-reviewed this paper.