acase@tulane.edu / 1 / Antonius Prader

No description available.

The study of acoustic resonance in fluid-filled piping systems with and without mean flow is important for the nuclear industry. For this industry, it is vital to understand the acoustic resonance in their systems; however, no comprehensive experimental benchmark data or accurate modeling tool exists for predicting such a phenomenon. The main goals of the current research are to create a new experimental data bank for the conditions not tested earlier using the configurations of straight lines and branches, and to evaluate the applicability of the linear wave solution using different damping methods and a computational fluid dynamic (CFD) code to simulate the acoustic resonance in fluid-filled piping systems. In this experimental study, data on resonant frequencies and resonant amplitudes are collected and analyzed for a frequency range of 20–500 Hz for straight and branched tubes by varying their wall thicknesses, materials, and branch configurations at different flow rates and outlet boundary conditions. To be closer to the nuclear industry medium, water is employed in our experiments, contrasting to the fact that most of the available experiments reported were with air at a much lower sonic velocity. I consider here, in particular, measurements at the end of closed branches, upstream, downstream, and at different locations of the main line, as well as the interactions of different sonic velocities along the main pipes. A small diameter is chosen for the branched experiments since the decrease in the width of the main line and the branches has a pronounced effect on the resonant amplitudes due to an increased interaction among the unsteady shear layers forming across the side branches. The experimental results show that there is a strong effect of turbulent flow, wall material, and wall thickness on resonant amplitudes at frequencies above ∼250 Hz. Numerical investigations are performed solving the one-dimensional (1D) linear wave equation with constant and frequency-dependent damping terms and a CFD code. Employing frequency-dependent damping methodologies shows improvement in terms of resonant amplitude prediction over constant volumetric drag method. Comparing the 1D and CFD results shows that the CFD solution yields better predictions. / February 2017

The subject of this thesis is the development of new ultrasound inspection techniques for detection of cracks that are smaller than the wavelength of the inspecting wave and the characterization of cracks in fluid-filled pipes as either surface-breaking or subsurface. The spectrum of the scattered field of a partially closed crack comprises harmonic components not expected to be found in the case of linear scatterers such as pores or inclusions. Paper A presents an experimental investigation into the linear reflection and generation of the 2nd harmonic component following the incidence of an ultrasonic wave onto a dry or water-confining interface formed by elasto-plastic steel-steel surfaces in contact. The results indicate that water has an unexpected effect on the reflection, at low interfacial pressures, suggesting that fluid mediated forces play a role not accounted for in current models. The level of the generation of the 2nd harmonic measured provides support for further development of the technique for detection of dry, partially closed cracks or fluid-filled, nearly open cracks. A theoretical model describing the nonlinear scattering of acoustic waves by surface-breaking cracks with faces in partial contact is presented in Paper B. Both linear and nonlinear response of the crack are shown to be the largest for a SV wave incident on the surface containing the crack at an angle just above the critical angle for longitudinal waves. A method which provides information on whether a fracture is surface-breaking or subsurface has been modelled and its optimal experimental set-up examined in Paper C. The main assumption of the model is that water carried by pressurized pipes infiltrates and fills a surface-breaking crack, while a subsurface crack is dry. The model simulates an inspection in which the modulation technique is employed and the surface hosting the crack is not accessible. A parameter, constructed with signals recorded in backscattering configuration during a modulation cycle, is examined and shown to provide a clear criterion to distinguish subsurface from surface-breaking cracks when a SV wave at 45 degree incidence is employed as a probe. Finally, in Paper D the modulation technique is experimentally tested on steel beams that host surface-breaking fatigue cracks. The method is shown to be a successful tool to distinguish a dry from a fluid-filled crack. Furthermore, it is revealed that the dynamics of the fluid needs to be accounted for in a more accurate simulation tool. / QC 20100906

The guided wave technique is commonly used for rapidly long-range pipeline inspection without removing the insulation of pipes. The torsional mode T(0,1) of the guided waves is usually generated to detect the defects in pipelines, since it has the advantage of being non-dispersive across the whole frequency range. However, a large number of pipelines are carrying fluid, wrapped with the coating material, and supported with clamp for the necessary manufacturing process in refinery and petro-chemical industrials. When these works are employed on the pipeline, the propagating guided waves may vary with the contents of material and how well the material compact on the pipe. Some energy of the incident guided wave in the pipe wall may leak into inside of contents or outside of wrapped materials and reduce the wave propagation distance. The effect of the fluid-filled pipe, the wrapped pipe, and the clamp support mounted on the pipe for guided wave propagation is investigated by both simulative and experimental methods. The wave structure of the T(0,1) mode in the pipes is analyzed by using the DISPERSE software for various cases to evaluate its influence to the guided wave propagation on the pipe. The amplitudes of the reflected signals from various features on the pipe are also measured using pipe screening system for calculating the attenuation of guided waves due to the features. The trend for the results is in good agreement between the experiments andpredictions for all cases of researches in this dissertation. It is found that the low viscosity liquid deposited in the pipe, such as water, diesel oil, and lubricant, has no effect on the torsional mode; while the high viscous of the fuel oil deposited in the pipe attenuates the reflection signal heavily for the pipe carrying fluid. In addition, both the full-filled and half-filled contents in the pipe are also studied in this case. The effects of the half-filled are the same as the full-filled results obtained. For the pipe wrapped with the coated material, the adhesive strength of the coated material is strong, such as bitumen and polyethylene; the attenuation of the guided waves is high; and there is almost no effect for mineral wool coating. Furthermore, the traveling distance of the guided waves in the pipe is also evaluated for various cases of the coated materials. The results indicate that the higher attenuation of the guided waves for the coated material, the shorter of the traveling distance in the pipe. For the clamp support mounted on pipe, the attenuation of the guided waves for the clamp support with a rubber gasket in between the pipe and the clamp is heavier than the case of clamp support without the rubber gasket is. Furthermore, the higher torque setting on the clamp (with or without the rubber gasket), the higher amplitude of the reflected signal is measured for the guided wave propagation. The effect of the frequency excitation is additionally demonstrated in this dissertation. It is noted that the higher amplitude of the reflected signal, the lower frequency excitation; moreover, theresonant effect is observed in the case of the clamp support with rubber gasket during the torque setting in the experiments. Good agreement has been obtained between the experiments and theoretical calculations of this effect.

In this work, we study the coupled dispersion characteristics of three distinct structural-acoustic waveguides, namely: -(1) a two-dimensional waveguide, (2) a fluid-filled circular cylindrical shell and (3)a fluid-filledelliptic cylindrical shell. Our primary interest is in finding coupled wavenumbers as functions of the fluid-structure coupling parameter(µ). Using the asymptotic solution methodology, we find the coupled wavenumbers as perturbations over the uncoupled wavenumbers of the component systems (the structure and the fluid). The asymptotic method provides us with analytical expressions of the coupled wavenumbers for small and large values of µ. The dispersion curves obtained from these extreme values of µ help in predicting the nature of the continuous transition of the wavenumber branches over the entire range of µ. Since the coupled wavenumbers are obtained as perturbations over the uncoupled wavenumbers, the perturbation term characterizes the effect of one medium over the other in terms of additional mass or stiffness. As is common in asymptotic methods, a particular form of the asymptotic expansion remains valid over a certain frequency range only. Hence, different scalings of the asymptotic parameter are used for different frequency ranges. In this regard, the method adopted uses principles of Matched Asymptotic Expansion (MAE). As mentioned above, we begin the study with a two-dimensional structural acoustic waveguide. Depending on the boundary condition at the top-edge of the fluid-layer (rigid or pressure-release), two cases are separately analyzed. In both these cases, only a single perturbation parameter (µ) is used. This is followed by the study of the axisymmetric mode vibration of a fluid-filled circular cylindrical shell. Here, in addition to , we include the Poisson’s ratio as another asymptotic parameter. The next problem studied is the beam mode (n =1)vibration of the same fluid-filled circular cylindrical shell. Here, the frequency is used as an asymptotic parameter (in addition to ) and the derivations proceed in two separate parts, one for the high frequency and the other for the low frequency. Having completed the n = 0 and n = 1 modes of the cylindrical shell, the higher order shell modes are studied using the simpler shallow shell theory. For the final system, viz., the elliptic cylindrical shell, another asymptotic parameter in the form of the eccentricity of the cross-section is used. Having derived the analytical expressions for the coupled wavenumbers and obtained the dispersion curves, a unified behavior of structural-acoustic systems is found to emerge. In all these systems, for small , the coupled wavenumbers are close to the in vacuo structural wavenumber and the wavenumbers of the rigid-walled acoustic duct. The measure of closeness is quantified by . As µ increases, these wavenumber branches get shifted continuously till for large µ, the coupled wavenumber branches are better identified as perturbations to the wavenumbers of the pressure-release acoustic duct. At the coincidence region, the coupled structural wavenumber branch transits to the coupled acoustic wavenumber branchand vice-versa. As a result, at coincidence frequencies, while the uncoupled wavenumber branches intersect, due to the coupling, there is no longer an intersection. These common characteristics are shared amongst all the systems despite the difference in geometries. This suggests that the above discussed features capture the essential physics of sound-structure coupling in waveguides.This workthus presents a novel unified view-point to the topic. Along the way, some additional novel studies are conducted which do contribute to the completeness of the work. The free wavenumbers determined from the asymptotic expressions are usedto calculate the forced response of the two-dimensional waveguide due to a δ forcing. Using this analysis, we are able to come up with a novel explanation of the observation that with coupling the dispersion curves cannot intersect. Additionally, the effect of bulk flow in the acoustic fluid is also comprehensively studied for the easier case of the two-dimensional waveguide. Further, the well-known universal dispersion relation for the higher order circumferential modes of the in vacuo circular cylindrical shell is re-derived using a simpler method.