Defects in underground pipes are detected by applying Gabor transforms on experimental guided wave signals and
comparing the experimental group velocity plots with the theoretical group velocity dispersion curves. Gabor transform,
which is a powerful signal processing tool, maps a signal into a two-dimensional space of time and frequency. Thus it
provides information about both when and at what frequency a signal arrives. Focus of this paper is to study the
applicability of cylindrical guided waves to detect defects in underground pipes using Gabor transform. Cylindrical
guided waves are generated by piezo-electric transducers. Guided waves are propagated through pipes that are buried in
the soil by placing transmitters on one end of the pipes and the receivers on the other end. The recorded signals are then
processed using 2-D Gabor Transform or Short Time Fourier Transform (STFT). Gabor transform converts the time-amplitude
signal into a time frequency signal which reveals the group velocities hidden in the signal. These
experimentally obtained group velocities are then compared with the theoretical velocities for cylindrical pipes
embedded in the soil. From the comparison of the theoretical and experimental group velocities, an effort has been made
to identify which modes are propagating through the embedded defective pipes and which modes are having difficulty to
propagate through the defective pipe wall. From this knowledge pipe wall defects can be detected.
This paper investigates if cylindrical guided waves can be effectively used for pipe wall defect detection in soil-embedded pipes. For this purpose guided waves are propagated through pipes that are buried in the soil by placing transmitters on one end of the pipes and the receivers on the other end. Received signals for both defect-free and defective pipes are subjected to wavelet transforms. It is found that when a Continuous Wavelet Transform (CWT) based algorithm is applied to analyze the received signals then it is very easy to make distinction between damaged and undamaged pipes. To investigate whether embedding the pipe in the soil makes it more difficult to detect the pipe wall defects, the same set of defective and defect-free pipes are analyzed before and after burying them in the soil. In both cases the defects are easily detected after analyzing the wavelet transformed signals. Interestingly it can be detected more easily for the buried pipes because the difference between the received signal strengths from defect-free and defective pipes is found to be greater for the buried pipes. For soil embedded pipes the ultrasonic energy scattered by the defect is absorbed by the surrounding soil making the energy reaching the receiver significantly weaker than that for the defect-free soil embedded pipe.
From various studies by different investigators it has been now well established that a number of cylindrical guided wave modes are sensitive to the pipe wall defects. Several investigations by these authors and other researchers showed that the strengths of the guided waves propagating through a pipe that is placed in air are reduced when the pipe wall defects are encountered. This reduction is expected because the pipe wall defects (gouge, dent, removed metal due to corrosion etc.) alter the pipe geometry, hampering the free propagation of guided wave modes. When water flows through the pipes, the guided wave technique becomes more challenging because the flowing water absorbs part of the propagating acoustic energy. Flowing water may also induce some standing modes. The propagating cylindrical guided wave modes become leaky modes in presence of the flowing water, in other words energy leaks into water. Therefore, the energy detected by a receiver, placed at a large distance from the transmitter, is reduced even for a defect free pipe. Further reduction in the signal strength occurs in presence of defects.
A recently developed semi-analytical technique called DPSM (Distributed Point Source Method) is improved and used to model the ultrasonic field in a fluid generated by an ultrasonic transducer and scattered by a solid plate of finite dimension. Earlier works on the ultrasonic field modeling by the DPSM technique have been limited to homogeneous fluids or nonhomogeneous media with infinite interfaces. This is the first attempt to model the complete ultrasonic field consisting of incident, reflected, transmitted and diffracted fields by a finite scatterer of any shape or size. No closed form analytical solution exists for ultrasonic field computation in presence of a scatterer and an ultrasonic transducer, both of which can have finite dimensions and any shape. Finite element solution for wave propagation analysis is very time consuming; hence, the semi analytical technique used here appears to be the method of choice for solving such practical problems. The paper shows how the scattered field varies as the acoustic properties and dimensions of the scatterer change.
DPSM (Distributed Point Source Method) is a computational technique that can be used to model the pressure field generated by ultrasonic acoustic transducers. This technique involves discretization of the transducer face of any geometrical shape, into a number of elemental surface areas. Point sources are placed at the centroids of the elemental surface areas. The strengths of the point sources are proportional to the surface areas. Pressure field at a given point is the cumulative effect of the pressure fields generated by all point sources. The accuracy of the computational technique depends on the sensor surface discretization. In this paper, circular transducers are modeled using the DPSM technique. This technique is applied to calculate the pressure field distribution in non- homogeneous fluids with interface. The non-homogeneous fluid is composed of two fluid half-spaced with the interface in front of the transducer face parallel or inclined to the transducer face. Pressure fields in both fluids for normal and angular incidence of the ultrasonic beam have been calculated using DPSM technique.