Glegg, Stewart A. L.

At the present time the Federal Highway Administration is developing a new highway noise prediction model called TNM that is anticipated to be a superior model than its predecessor STAMINA 2.0. TNM will use frequency dependent measured noise source heights as a function of speed, acceleration, pavement type, road grade and vehicle class in its propagation loss and insertion loss calculations. Also, TNM will split the equivalent source height into two sub-sources at specified heights and calculate the insertion loss of noise barriers from these two sub-sources. The accuracy of this concept is investigated in this thesis, and it is found that on the average the two sub-source approach is less accurate than the single equivalent source approximation but will overestimate the attenuated levels, thus a conservative barrier height can be designed from this result.

Broadband self-noise generated by rotating blades in a subsonic ducted propfan is studied for a hard walled cylindrical duct in a uniform flow. An expression for the induct sound power radiated by three self-noise mechanisms is derived: the Turbulent-Boundary-Layer-Trailing-Edge noise, the Laminar-Boundary-Layer-Vortex-Shedding noise and the Trailing-Edge-Bluntness noise. The present theory uses NASA's self-noise prediction methodology for an isolated airfoil. An efficient method of programming is presented which reduces the time of computation for multiple radial modes. The results obtained are presented, discussed and compared with Blade-Tip-Boundary-Layer fan noise predictions obtained using the SDPF code developed at FAU. The most important parameters which affect self-noise are found to be the angle of attack, the effective Mach number and the chord length of the blade. For high angles of attack, the TBL-TE noise gives significant amount of sound power especially at the low frequencies. For low effective Mach numbers and at certain angles of attack, the LBL-VS noise can have high power levels in the mid and high frequencies. Trailing edge bluntness noise appeared to give insignificant amounts of energy over the whole spectrum compared to the other self-noise mechanisms.

A theoretical model has been developed to compute the vertical array directional response for surface generated ambient noise in a shallow water environment. The cross spectrum function is based on a normal mode solution to the wave equation in which the effective depth approximation is used to yield closed form solutions for two distinct mode types. The effective depth modes encompass the shallowest grazing angles where the bottom reacts as a pressure release surface to the incident plane waves. The rigid bottom solution takes over as the grazing angle increases and attenuation becomes significant. The computed vertical array beam output was compared to other models including a fast field wavenumber integration method and a multipath eigenray method with mixed results. The results indicated good agreement for both comparisons with the realization that the effective depth model is sensitive to the approximation discontinuity at the mode transition point.

Buckingham et al. (Nature Vol. 356, p 327) first introduced the concept of acoustic imaging using ambient noise as a method for passively detecting objects in the ocean. Several analytical studies followed, and it was shown that a two dimensional acoustic image could be obtained based on this approach, and that at least 900 pixels are necessary to restitute the details of spherical objects placed in an underwater sound channel. The alternative approach described in this paper is based on a signal processing which uses the broadband nature of the ambient noise in the ocean, and therefore, optimizes the use of available sound energy scattered by the object. Images with thousands of pixels can be obtained using a relatively small number of transducers. This method has been validated using simple experiments in air, scaled to represent an ocean application, and results showing images of various objects will be presented.

Broadband noise generated by subsonic ducted propfans is studied for a hard walled flanged duct with an infinite centerbody and in a uniform flow. The interaction between boundary layers and blades and the impingement of the blade wakes on the stators are the main sources of noise considered in the forward and rear arc respectively. An efficient method of programming is presented which reduces the time of computation of expressions. The results obtained are presented and discussed. The boundary layer noise is found to have no sufficient energy at low frequencies regarding the available data, unless the boundary layer thickness is increased by a small amount. Finally, good predictions are obtained with the wake noise although some parameters would require a better evaluation.

This study proposed to determine an ocean surface bubble radius distribution using theoretical noise models, experimental noise spectral data and a theoretical model for the bubble size distribution. The obtained distribution was compared to previous experimental distributions. A good agreement was found for the radius upper cut-off but the shape and the peak radius of the distribution are different from experimental observations. The most probable reasons for these differences were given. Moreover, the validity of recent theoretical bubble noise models was checked. Eventually, the main limitations of the study were pointed out and directions for future works were given.

Laboratory scale measurements were conducted to validate numerical prediction models used to predict the acoustic field in a shallow water ocean environment. Experimental measurements were conducted in a range independent environment which included the effects of shear in marine sediments and in a depth varying range dependent environment. Good agreement between the experimental measurements and the numerical prediction codes were obtained using optimized values for the input parameters of the environmental model. In comparing experimental measurements to the numerical prediction codes it became apparent that the codes were very sensitive to the input parameters describing the bottom boundary of the ocean waveguide.

A theoretical model accounting for underwater ambient noise due to whitecap-generated bubbles and drop-impacts is developed to investigate the possible mechanisms of ocean noise generation. A general analysis is laid down, leading to expressions of the noise spectrum in an undefined environment. Subsequently the cases of isovelocity and stratified deep oceans are considered, and specific expressions are derived. The stratified deep ocean estimations of the directional spectrum are then compared to experimental data and a good agreement is found. Finally the absolute noise levels due to bubbles and drop impacts are discussed, and bubble noise is found to be larger than drop impact noise. The collective bubble oscillation hypothesis is also discussed, and although clues are given for and against this hypothesis no rigorous justification for this has been found.

A scaled, horizontally stratified shallow water marine acoustic model is constructed for the purpose of investigating the sound field resulting from long range propagation. The characteristics of this sound field in the water column are strongly dependent upon properties of the surficial sediment. One effect is the conversion from compressional waves in the water column to shear waves in the sediment at grazing angles less than the critical. To model a shallow marine environment, concrete is used as a substrate rock and laminating epoxy is used to model a surficial sediment. Preliminary tests of the effects of the model's environment are performed, and the range dependent sound field as a function of depth is measured for several CW frequencies. These sound field profiles are compared with an approximate predictive theory, and with a numerical solution which treats the sediment properties exactly.

The interaction between vibrating structures and fluids can have a profound influence upon the natural frequencies of the structure's vibration. This study examines one specific structure; a thin circular plate with the rarely studied free edge condition. It starts by considering a completely free plate in a vacuum and then, using receptance matching, utilises this result to determine the effects, on the natural frequencies, of a centrally located driving rod. Then, using the same technique, a result for the drive admittance of the fluid loaded plate is adapted to predict the natural frequencies of the same structure when subjected to significant fluid loading. All these results are then compared to those obtained from experiments.