Nondestructive Evaluation of the Depth of Cracks in Concrete Plates Using Surface Waves
MetadataShow full item record
Concrete structures can often be modeled as plates, for example, bridges, tunnel walls and pipes. Near-surface damage in concrete structures mostly takes the form of cracking. Surface-breaking cracks affect concrete properties and structural integrity; therefore, the nondestructive evaluation of crack depth is important for structural monitoring, strengthening and rehabilitation. On the other hand, material damping is a fundamental parameter for the dynamic analysis of material specimens and structures. Monitoring damping changes is useful for the assessment of material conditions and structural deterioration. The main objective of this research is to develop new methodologies for depth evaluation of surface-breaking cracks and the evaluation of damping in concrete plates. Nondestructive techniques based on wave propagation are useful because they are non-intrusive, efficient and cost effective. Previous studies for the depth evaluation of surface-breaking cracks in concrete have used diffracted compressional waves (P-waves). However, surface waves exhibit better properties for the characterization of near surface defects, because (a) surface waves dominate the surface response, they carry 67% of the wave propagation energy, and present lower geometrical attenuation because the propagating wave front is cylindrical; and (b) the penetration depth of Rayleigh waves (R-waves) depends on their frequency. Most of the R-wave energy concentrates at a depth of one-third of their wavelengths. The transmission of R-waves through a surface-breaking crack depends on the crack depth; this depth sensitivity is the basis for the so-called Fourier transmission coefficient (FTC) method. R-waves only exist in a half-space (one traction-free surface); whereas in the case of a plate (two traction-free surfaces), Lamb modes are generated. Fundamental Lamb modes behave like R-waves at high frequencies, because their wavelengths are small relative to the plate thickness. Lamb modes are not considered in the standard FTC method, and the FTC method is also affected by the selected spacing between receivers. The FTC calculation requires the use of an explicit time window for the identification of the arrival of surface waves, and the selection of a reliable frequency range. This research presents theoretical, numerical and experimental results. Theoretical aspects of Lamb modes are discussed, and a theoretical transfer function is derived, which can be used to study changes of Lamb modes in the time and frequency domains as a function of distance. The maximum amplitude of the wavelet transform varies with distance because of the dispersion of Lamb modes and the participation of higher Lamb modes in the response. Numerical simulations are conducted to study the wave propagation of Lamb modes through a surface-breaking crack with different depths. The surface response is found to be dominated by the fundamental Lamb mode. Using the 2D Fourier transform, the incident, transmitted and reflected fundamental Lamb modes are extracted. A transmission ratio between the transmitted and incident modes is calculated, which is sensitive to crack depths (d) normalized to the wavelength (λ) in a range (d / λ) = 0.1 to 1/3. A new wavelet transmission coefficient (WTC) method for the depth evaluation of surface-breaking cracks in concrete is proposed to overcome the main limitations of the FTC method. The WTC method gives a global coefficient that is correlated with the crack depth, which does not require time windowing and the pre-selection of a frequency bandwidth. To reduce the effects of wave reflections, which are present in the FTC method because of the non-equal spacing configuration, a new equal spacing configuration is used in the WTC method. The effects of Lamb mode dispersion are also reduced. In laboratory tests, an ultrasonic transmitter with central frequency at 50kHz is used as a source; the 50kHz frequency is appropriate for the concrete plate tested (thickness 80mm), because the fundamental Lamb modes have converged to the Rayleigh wave mode. The new method has also been used in-situ at Hanson Pipe and Precast Inc., Cambridge, Ontario, Canada, and it shows potential for practical applications. In general, the evaluation of material damping is more difficult than the measurement of wave velocity; the dynamic response and attenuation of structural vibrations are predominantly controlled by damping, and the damping is typically evaluated using the modal analysis technique, which requires considerable efforts. The existing methods based on surface waves, use the Fourier transform to measure material damping; however, an explicit time window is required for the spectral ratio method to extract the arrival of surface wave; in addition, a slope of the spectral ratio varies for different frequency ranges, and thus a reliable frequency range needs to be determined. This research uses the wavelet transform to measure material damping in plates, where neither an explicit time window nor the pre-selection of a frequency bandwidth are required. The measured material damping represents an average damping for a frequency range determined by source. Both numerical and experimental results show good agreement and the potential for practical applications.