Periodic inspections of aging concrete structures are useful for identifying concrete degradation before it becomes a serious problem. Where visual inspection is impractical or impossible, nondestructive evaluation methods can be used. Three such methods – spectral analysis of surface waves, ground penetrating radar, and acoustic tomography – are particularly useful in dam and spillway inspections.
By Dennis A. Sack, Larry D. Olson, and Hunter A. Yarbrough
As large dams age, it becomes increasingly important to determine the condition of their concrete elements and to track changes in this condition over time. The elements that are of the most concern, because they are the most critical structural elements of the dam, include the spillway concrete, spillway subgrade, dam wall concrete, and interiors of thin arch dams.
Because of the large mass of concrete in dams and spillways and the long expected life of these structures, they are susceptible to degradation mechanisms that can start as minor problems and be present for years. These mechanisms include freeze-thaw damage on the downstream face and crown, seepage under and around outflow pipes and spillways, slow-developing cracks in the dam interior, and erosion due to water flow and weathering.
The presence of initial degradation tends to accelerate future problems. For example, surfaces damaged by the freeze-thaw cycle tend to hold more moisture than undamaged surfaces, thus leading to greater damage during future freeze-thaw cycles. In addition, small cracks in the dam face become stress concentrators that lead to deeper cracking. Cracks that seep can lead to erosion and cracking as a result of the action of the water. And seepage under a spillway or around an outflow pipe slowly erodes the subgrade, leading to faster flow rates and even greater erosion.
Periodic inspection is a dam owner’s best defense against these threats. However, visual inspections often only reveal problems that have developed into major degradation. In addition, some types of damage – such as voids under a spillway slab or cracking in the dam interior – are difficult or impossible to uncover using only visual means.
This is where nondestructive evaluation (NDE) techniques enter the picture. NDE methods use sound waves, radio waves, and other types of low-level energy to penetrate the concrete. These methods can be used to locate a void under a spillway; measure the depth and severity of freeze-thaw or other surface damage; and create an image slice through the interior of a dam to show cracks, debonded joints, and other degraded zones. Most of these evaluations can be performed without dewatering the dam, and all are completely nondestructive and nonintrusive.
Nondestructive evaluation techniques
Many NDE methods can be used to investigate the condition of concrete elements in a dam. These methods include impact echo, slab impulse response, spectral analysis of surface waves (SASW), ground penetrating radar (GPR), and acoustic tomography (AT).
In particular, three techniques are most useful in dam investigations. The first is SASW, which is used to measure the condition of concrete vs. depth from the test surface, as well as to measure the depth of cracks.1 The second is GPR, which is used to map out voids and water seepage paths under a spillway, as well as to provide information on the spacing and depth of any reinforcing in the concrete. The third is AT, which uses sound wave transmission through a thin-arch dam to create a tomographic image slice from the upstream to downstream faces.
Spectral analysis of surface waves
The SASW method is based on measurements of surface waves propagating in layered elastic media. The ratio of surface wave velocity to shear wave velocity varies with Poisson’s ratio. However, reasonable estimates of Poisson’s ratio and the mass density of concrete and other materials can be made with only a small effect on the accuracy of the shear wave velocity profile. By using the shear wave velocity and reasonable estimates of the mass density of the material layers, calculations can be made of shear moduli for low-strain amplitudes.
Surface wave (also termed Rayleigh or R-wave) velocity varies with frequency in a layered system. This variation is called dispersion. A plot of surface wave velocity vs. wavelength is called a dispersion curve.
The SASW tests and analyses generally are performed in three phases:
- Collection of data in-situ;
- Construction of an experimental dispersion curve using field data; and
- Forward modeling of the experimental dispersion curve, if desired, to match other theoretical curves so that a profile of shear wave velocity vs. depth can be constructed.
Wavelength, frequency, and wave velocity are related as follows:
- Vr = f x l
- Vr is the wave velocity;
- f is the frequency; and
- l is the wavelength.
Surface wave dispersion can be expressed using a plot of surface wave velocity vs. wavelength, called a dispersion curve plot. For sound concrete, the dispersion curve plot is flat (constant velocity) vs. wavelength throughout the concrete thickness. If degradation or cracking is present, the dispersion curve will have zones of lower velocity. The depth of the crack or degradation can be estimated from the wavelength range of the dispersion curve.
SASW field tests are conducted using a pair of vibration transducers. The transducers are typically displacement transducers or accelerometers mounted to or held in contact with the concrete surface. Impacts are made with a small hammer in line with the receivers. A computerized signal conditioning and data acquisition system collects the waveforms from the receivers and records the signals on a computer hard drive for spectral (frequency) analyses. The phase information of the transfer function (cross power spectrum) between the two receivers for each frequency is the key spectral measurement.
The experimental dispersion curve is developed using the phase data from a given site, by knowing the phase at a given frequency and then calculating the travel time between receivers of that frequency and/or wavelength by:
- t = l / 360 x f
- t is the travel time between receivers at a given frequency and/or wavelength;
- l is the wavelength; and
- f is the frequency.
Surface wave velocity is obtained by dividing the receiver spacing by the travel time at a frequency:
- Vr = X / t
- Vr is the wave velocity;
- X is the receiver spacing; and
- t is the travel time between receivers at a given frequency.
Repeating the above procedure for any given frequency allows for evaluation of the surface wave velocity corresponding to a given wavelength and construction of the dispersion curve.
The phase data typically is viewed in the field on the personal computer-based data acquisition system, to ensure acceptable data is being collected. The phase data is then sent to the office for further processing. The phase of the cross power spectrum (transfer function) between the two receivers and the coherence function are used to create the dispersion curves. After frequency masking of the phase record pair from each test location, an experimental field dispersion curve is developed that is the plot of surface wave velocity vs. wavelength.
Figure 1 shows experimental dispersion curve data from a test at a sound location on a dam face. Visual inspection of this location revealed no evidence of surface degradation, and SASW data results showed no evidence of degradation. Thus, this plot shows the relatively flat, high-velocity (6,500 feet per second, fps) dispersion curve typical of sound concrete.
Figure 2 shows a sample data plot from a point farther down the same dam face, where the surface appeared to be distressed from freeze-thaw damage. SASW test results support the degradation of the concrete at the surface and also provide an estimate as to the depth of degradation into the concrete face. The degradation is evidenced by the low initial velocity of the plot (3,000 fps), which begins to increase at a wavelength of about 0.3 foot. This is typical of concrete with significant surface degradation to a depth of about 0.3 foot, at which point the sound concrete at deeper depths begins to raise the average surface wave velocity for longer wavelengths.
Ground penetrating radar
The GPR method involves moving an antenna across a test surface while periodically pulsing the antenna and recording the received echoes. Pulses are sent out from the GPR computer driving the antenna, at a frequency range centered on the design center frequency of the antenna. Antenna center frequencies used in NDE investigations vary widely, depending on the structural geometry and the information desired. To locate rebar or detect voids under thin slab-type spillways, a 900 to 1,500 megaHertz (MHz) antenna typically is used.2 For thicker slabs or other thick concrete elements, antennae with a center frequency of 400 MHz or lower often are selected. Lower-frequency antennae allow for deeper penetration, but at the sacrifice of resolution.
The electromagnetic wave pulses propagate through the material directly under the antenna. Some energy reflects back when the wave encounters a change in electrical impedance, such as at rebar or an air-filled void. The antenna receives these echoes, which are amplified and filtered in the GPR computer, then digitized and stored. A distance wheel records scan distance across the test surface, and embedded features can be located as a given distance from the scan start position. For repetitive scanning, a standard survey is designed and adhered to as field conditions allow to minimize mistakes and maximize data quality.
Typical scan intervals for tests on thinner elements such as spillway slabs are at lateral intervals of about 30 pulses per foot. The resulting raw data is in the form of echo amplitude vs. time. By inputting the dielectric constant (values of 5.5 to 6 are typical of older concrete) and estimating the signal zero point, the echo time data can be converted to echo depth. If more accurate depth data is required, a depth calibration can be done if an embedment of a known depth is available to scan over. The scans are then translated into waterfall plots of all the individual data traces collected, with the lightness or darkness (or color) of each point being set by the amplitude and polarity (positive or negative) of the data at a given depth in each trace.
A nondestructive investigation recently was conducted on the spillway slab of a small dam in the southern U.S. This 50-foot-wide concrete slab had unknown reinforcing and showed evidence of settlement and distortion. In addition, there were small water flows near the base and sides of the spillway that were presumed to be a result of flow under the spillway slab. The owner requested the NDE to examine the concrete slab and its subgrade support conditions. Objectives of this investigation were to locate any reinforcing and to map out voids under the slab for subsequent mitigation. As part of this investigation, a set of GPR scans was performed across the spillway, at 5-foot intervals down the length of the structure. These scans showed clear evidence of water-filled voids under the spillway. They also showed the presence of rebar in the slab concrete, at nominal 12-inch centers.
Figure 3: A ground penetrating radar scan across a spillway slab revealed the presence of rebar in the concrete, as well as a water- or air-filled void under the slab.
Figure 3 shows a GPR scan from 65 feet down the spillway slab. It is easy to see the periodic hyperbolic reflectors spaced along the top of the scan. These are from the rebar in the slab, at a spacing of about 1 foot center to center. There also is a clear zone of stronger reflections from the subgrade and from the slab bottom. This zone is located near the left side of the plot. These stronger reflections are typical of a water- or air-filled void under the slab, with both the bottom of concrete echo (right below the rebar) and the top of soil echo becoming much stronger than normal.
The AT method is used to provide information on the interior structure and condition of concrete elements where access is available on two sides of the element. Sound wave energy is sent from a number of evenly spaced source locations on one face to a number of evenly spaced receiver locations on the opposite face.3 With enough source-receiver combinations, the resulting data set can be processed with a tomographic processing software package, creating a velocity contour image of the interior of the element.
The application of the method to thin-arch concrete dams has been refined to improve data collection rates by using a hydrophone string suspended in the water on the upstream face as a receiver array. This allows collection of many ray paths of data from each source impact location. With enough receivers, the full tomographic data set can be collected in a single top-to-bottom pass of the sound wave source. The full set of data paths, or rays, ideally is a dense grid of overlapping lines.
Typical sound wave sources for AT tests on dams have included instrumented hammers used by rope-suspended climbers, or electrically-operated solenoid impactors mounted on a rolling frame lowered down the downstream face. In a recent investigation, a technician on a rope used a 3-pound instrumented modal hammer to impact the dam face. The impacts from the hammer allowed the input of high-amplitude signals that could be picked up by the hydrophone string even at high source-to-receiver angles and distances and in areas with poor surface conditions.
The data collected from AT tests are analyzed to pick the compression wave arrival times. Compression wave travel times and geometry are then input into a tomography calculation software package that computes the velocities in the interior of the dam based on all the signals that pass through each interior point. The computed velocity data is then input into an advanced modeling software to produce velocity tomogram images. (See Figure 4.)
Figure 4: This acoustic tomogram of a thin arch concrete dam shows several areas of low velocity that correspond to low strength/ degraded concrete areas, primarily on the downstream face of the dam, at left.
Typically, tomographic images are analyzed to look at the velocity changes within the concrete. Areas with lower velocity correspond to weaker, less dense concrete, while those with higher velocities are considered to be sound concrete. The results also can show areas with cracking damage or other discontinuities.
In tomographic analysis, it also is possible to get “artifacts” in the data in areas with low ray density (not many data paths passing through an area) or in areas with non-linear or non-homogeneous velocities. The tomogram in Figure 4 includes a few of these artifacts at or near the upstream face. The artifacts were most severe near the top of the dam, where the ray density was very low. Close examination of the data showed most of these zones are localized right at the underwater hydrophone receiver locations. These artifacts likely are due to the effects of variable coupling between the receivers and the concrete through a rubberized membrane and drainage fabric layer on the upstream face.
The AT test results in Figure 4 provide a picture of the interior condition of the concrete through a slice of the dam from the downstream to upstream faces. These results show generally sound concrete through the dam cross section, with only minor downstream face surface degradation, down to about elevation 85 feet. Below an elevation of 85 feet, the degradation becomes more uniform, with the AT results showing degraded concrete out to a depth of about 5 to 7 feet or more from the downstream face. As seen in the figure, the most severe areas of degradation from the AT results are at elevations 22 to 34 feet and 44 to 57 feet.
Each of the three NDE methods presented can provide information on concrete damage and degradation not available from a simple visual inspection. The time required for a typical NDE investigation varies based on the size and geometry of the structure and the nature of the problems to be investigated. The majority of investigations are conducted during site visits of one to five days, at a cost of less than $30,000. Together, these methods provide a powerful and economical set of tools that dam owners and operators can use to ensure the structures in their care continue to safely perform.
When SASW, GPR, and AT testing are combined and analyzed in conjunction, they provide the information essential for assessing the dam’s longevity and structural safety.
1Olson, Larry D., Y. Fai Chan, Dennis A. Sack, Michael F. Dumont, Robert T. Gilmore, and John T. Christy, “The Use of Impact-Echo and Spectral-Analysis-of-Surface-Wave Methods for the Concrete Investigation of the Rogers Dam Spillway Structure,” Proceedings of the 56th American Power Conference, Illinois Institute of Technology, Chicago, Ill., 1994.
2Hollema, David, and Larry D. Olson, “Application of a Combined Nondestructive Evaluation Approach to Detecting Subgrade Voids Below a Dam Spillway,” Proceedings of the Symposium on the Application of Geophysics to Engineering and Environmental Problems, Environmental and Engineering Geophysical Society, Denver, Colo., 2004.
3Billington, Edward D., Dennis A. Sack, and Larry D. Olson, “Sonic Pulse Velocity Testing to Assess Condition of a Concrete Dam,” Proceedings of the Symposium on the Application of Geophysics to Engineering and Environmental Problems, Environmental and Engineering Geophysical Society, Denver, Colo., 2001.
Dennis Sack, vice president and associate engineer with Olson Engineering, designs and develops instrumentation to test the integrity of deep foundations, dams, and tunnels. Larry Olson, P.E., president and principal engineer with Olson Engineering, has managed and performed nondestructive testing and evaluation since 1985. Hunter Yarbrough, geophysical project engineer with Olson Engineering, has participated in many nondestructive evaluations of deep foundations, dams, and levee systems.